Show of Hands:Votes: Party classification:: Votes_RFEX_cnk04_rest
bdanalytics
Date: (Sun) Jun 05, 2016
Introduction:
Data: Source: Training: “https://inclass.kaggle.com/c/can-we-predict-voting-outcomes/download/train2016.csv”
New: “https://inclass.kaggle.com/c/can-we-predict-voting-outcomes/download/test2016.csv”
Time period:
Synopsis:
Based on analysis utilizing <> techniques,
Summary of key steps & error improvement stats:
Prediction Accuracy Enhancement Options:
- transform.data chunk:
- derive features from multiple features
- manage.missing.data chunk:
- Not fill missing vars
- Fill missing numerics with a different algorithm
- Fill missing chars with data based on clusters
[](.png)
Potential next steps include:
- Organization:
- Categorize by chunk
- Priority criteria:
- Ease of change
- Impacts report
- Cleans innards
- Bug report
- all chunks:
- at chunk-end rm(!glb_)
- manage.missing.data chunk:
- cleaner way to manage re-splitting of training vs. new entity
- extract.features chunk:
- Add n-grams for glbFeatsText
- “RTextTools”, “tau”, “RWeka”, and “textcat” packages
- Add n-grams for glbFeatsText
- fit.models chunk:
- Classification: Plot AUC Curves for all models & highlight glbMdlSel
- Prediction accuracy scatter graph:
- Add tiles (raw vs. PCA)
- Use shiny for drop-down of “important” features
Use plot.ly for interactive plots ?
- Change .fit suffix of model metrics to .mdl if it’s data independent (e.g. AIC, Adj.R.Squared - is it truly data independent ?, etc.)
- create a custom model for rpart that has minbucket as a tuning parameter
varImp for randomForest crashes in caret version:6.0.41 -> submit bug report
- Probability handling for multinomials vs. desired binomial outcome
- ROCR currently supports only evaluation of binary classification tasks (version 1.0.7)
extensions toward multiclass classification are scheduled for the next release
- fit.all.training chunk:
- myplot_prediction_classification: displays ‘x’ instead of ‘+’ when there are no prediction errors
- Compare glb_sel_mdl vs. glb_fin_mdl:
- varImp
- Prediction differences (shd be minimal ?)
- Move glb_analytics_diag_plots to mydsutils.R: (+) Easier to debug (-) Too many glb vars used
- Add print(ggplot.petrinet(glb_analytics_pn) + coord_flip()) at the end of every major chunk
- Parameterize glb_analytics_pn
- Move glb_impute_missing_data to mydsutils.R: (-) Too many glb vars used; glb_<>_df reassigned
- Do non-glm methods handle interaction terms ?
- f-score computation for classifiers should be summation across outcomes (not just the desired one ?)
- Add accuracy computation to glb_dmy_mdl in predict.data.new chunk
- Why does splitting fit.data.training.all chunk into separate chunks add an overhead of ~30 secs ? It’s not rbind b/c other chunks have lower elapsed time. Is it the number of plots ?
- Incorporate code chunks in print_sessionInfo
- Test against
- projects in github.com/bdanalytics
- lectures in jhu-datascience track
Analysis:
rm(list = ls())
set.seed(12345)
options(stringsAsFactors = FALSE)
source("~/Dropbox/datascience/R/mycaret.R")
source("~/Dropbox/datascience/R/mypetrinet.R")
source("~/Dropbox/datascience/R/myplclust.R")
source("~/Dropbox/datascience/R/myplot.R")
source("~/Dropbox/datascience/R/myscript.R")
source("~/Dropbox/datascience/R/mytm.R")
if (is.null(knitr::opts_current$get(name = 'label'))) # Running in IDE
debugSource("~/Dropbox/datascience/R/mydsutils.R") else
source("~/Dropbox/datascience/R/mydsutils.R") ## Loading required package: caret## Loading required package: lattice# Gather all package requirements here
suppressPackageStartupMessages(require(doMC))
glbCores <- 6 # of cores on machine - 2
registerDoMC(glbCores)
suppressPackageStartupMessages(require(caret))
require(plyr)## Loading required package: plyrrequire(dplyr)## Loading required package: dplyr##
## Attaching package: 'dplyr'## The following objects are masked from 'package:plyr':
##
## arrange, count, desc, failwith, id, mutate, rename, summarise,
## summarize## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, unionrequire(knitr)## Loading required package: knitrrequire(stringr)## Loading required package: stringr#source("dbgcaret.R")
#packageVersion("snow")
#require(sos); findFn("cosine", maxPages=2, sortby="MaxScore")
# Analysis control global variables
# Inputs
# url/name = "<PathPointer>"; if url specifies a zip file, name = "<filename>";
# or named collection of <PathPointer>s
# sep = choose from c(NULL, "\t")
glbObsTrnFile <- list(url = "https://inclass.kaggle.com/c/can-we-predict-voting-outcomes/download/train2016.csv"
# or list(url = c(NULL, <.inp1> = "<path1>", <.inp2> = "<path2>"))
#, splitSpecs = list(method = "copy" # default when glbObsNewFile is NULL
# select from c("copy", NULL ???, "condition", "sample", )
# ,nRatio = 0.3 # > 0 && < 1 if method == "sample"
# ,seed = 123 # any integer or glbObsTrnPartitionSeed if method == "sample"
# ,condition = # or 'is.na(<var>)'; '<var> <condition_operator> <value>'
# )
)
glbObsNewFile <- list(url = "https://inclass.kaggle.com/c/can-we-predict-voting-outcomes/download/test2016.csv")
glbObsDropCondition <- NULL # : default
# enclose in single-quotes b/c condition might include double qoutes
# use | & ; NOT || &&
# '<condition>'
# 'grepl("^First Draft Video:", glbObsAll$Headline)'
# 'is.na(glbObsAll[, glb_rsp_var_raw])'
# '(is.na(glbObsAll[, glb_rsp_var_raw]) & grepl("Train", glbObsAll[, glbFeatsId]))'
# 'is.na(strptime(glbObsAll[, "Date"], glbFeatsDateTime[["Date"]]["format"], tz = glbFeatsDateTime[["Date"]]["timezone"]))'
#nrow(do.call("subset",list(glbObsAll, parse(text=paste0("!(", glbObsDropCondition, ")")))))
glb_obs_repartition_train_condition <- NULL # : default
# "<condition>"
glb_max_fitobs <- NULL # or any integer
glbObsTrnPartitionSeed <- 123 # or any integer
glb_is_regression <- FALSE; glb_is_classification <- !glb_is_regression;
glb_is_binomial <- TRUE # or TRUE or FALSE
glb_rsp_var_raw <- "Party"
# for classification, the response variable has to be a factor
glb_rsp_var <- "Party.fctr"
# if the response factor is based on numbers/logicals e.g (0/1 OR TRUE/FALSE vs. "A"/"B"),
# or contains spaces (e.g. "Not in Labor Force")
# caret predict(..., type="prob") crashes
glb_map_rsp_raw_to_var <- #NULL
function(raw) {
# return(raw ^ 0.5)
# return(log(raw))
# return(log(1 + raw))
# return(log10(raw))
# return(exp(-raw / 2))
#
# chk ref value against frequencies vs. alpha sort order
ret_vals <- rep_len(NA, length(raw)); ret_vals[!is.na(raw)] <- ifelse(raw[!is.na(raw)] == "Republican", "R", "D"); return(relevel(as.factor(ret_vals), ref = "D"))
# as.factor(paste0("B", raw))
# as.factor(gsub(" ", "\\.", raw))
}
#if glb_rsp_var_raw is numeric:
#print(summary(glbObsAll[, glb_rsp_var_raw]))
#glb_map_rsp_raw_to_var(tst <- c(NA, as.numeric(summary(glbObsAll[, glb_rsp_var_raw]))))
#if glb_rsp_var_raw is character:
#print(table(glbObsAll[, glb_rsp_var_raw], useNA = "ifany"))
# print(table(glb_map_rsp_raw_to_var(tst <- glbObsAll[, glb_rsp_var_raw]), useNA = "ifany"))
glb_map_rsp_var_to_raw <- #NULL
function(var) {
# return(var ^ 2.0)
# return(exp(var))
# return(10 ^ var)
# return(-log(var) * 2)
# as.numeric(var)
# levels(var)[as.numeric(var)]
sapply(levels(var)[as.numeric(var)], function(elm)
if (is.na(elm)) return(elm) else
if (elm == 'R') return("Republican") else
if (elm == 'D') return("Democrat") else
stop("glb_map_rsp_var_to_raw: unexpected value: ", elm)
)
# gsub("\\.", " ", levels(var)[as.numeric(var)])
# c("<=50K", " >50K")[as.numeric(var)]
# c(FALSE, TRUE)[as.numeric(var)]
}
# print(table(glb_map_rsp_var_to_raw(glb_map_rsp_raw_to_var(tst)), useNA = "ifany"))
if ((glb_rsp_var != glb_rsp_var_raw) && is.null(glb_map_rsp_raw_to_var))
stop("glb_map_rsp_raw_to_var function expected")
# List info gathered for various columns
# <col_name>: <description>; <notes>
# USER_ID - an anonymous id unique to a given user
# YOB - the year of birth of the user
# Gender - the gender of the user, either Male or Female
# Income - the household income of the user. Either not provided, or one of "under $25,000", "$25,001 - $50,000", "$50,000 - $74,999", "$75,000 - $100,000", "$100,001 - $150,000", or "over $150,000".
# HouseholdStatus - the household status of the user. Either not provided, or one of "Domestic Partners (no kids)", "Domestic Partners (w/kids)", "Married (no kids)", "Married (w/kids)", "Single (no kids)", or "Single (w/kids)".
# EducationalLevel - the education level of the user. Either not provided, or one of "Current K-12", "High School Diploma", "Current Undergraduate", "Associate's Degree", "Bachelor's Degree", "Master's Degree", or "Doctoral Degree".
# Party - the political party for whom the user intends to vote for. Either "Democrat" or "Republican
# Q124742, Q124122, . . . , Q96024 - 101 different questions that the users were asked on Show of Hands. If the user didn't answer the question, there is a blank. For information about the question text and possible answers, see the file Questions.pdf.
# currently does not handle more than 1 column; consider concatenating multiple columns
# If glbFeatsId == NULL, ".rownames <- as.numeric(row.names())" is the default
glbFeatsId <- "USER_ID" # choose from c(NULL : default, "<id_feat>")
# glbFeatsCategory <- "Hhold.fctr" # choose from c(NULL : default, "<category_feat>")
glbFeatsCategory <- "Q109244.fctr" # choose from c(NULL : default, "<category_feat>") -> OOB performed worse than "Hhold.fctr"
# User-specified exclusions
glbFeatsExclude <- c(NULL
# Feats that shd be excluded due to known causation by prediction variable
# , "<feat1", "<feat2>"
# Feats that are factors with unique values (as % of nObs) > 49 (empirically derived)
# Feats that are linear combinations (alias in glm)
# Feature-engineering phase -> start by excluding all features except id & category &
# work each one in
, "USER_ID", "YOB", "Gender", "Income", "HouseholdStatus", "EducationLevel"
,"Q124742","Q124122"
,"Q123621","Q123464"
,"Q122771","Q122770","Q122769","Q122120"
,"Q121700","Q121699","Q121011"
,"Q120978","Q120650","Q120472","Q120379","Q120194","Q120014","Q120012"
,"Q119851","Q119650","Q119334"
,"Q118892","Q118237","Q118233","Q118232","Q118117"
,"Q117193","Q117186"
,"Q116797","Q116881","Q116953","Q116601","Q116441","Q116448","Q116197"
,"Q115602","Q115777","Q115610","Q115611","Q115899","Q115390","Q115195"
,"Q114961","Q114748","Q114517","Q114386","Q114152"
,"Q113992","Q113583","Q113584","Q113181"
,"Q112478","Q112512","Q112270"
,"Q111848","Q111580","Q111220"
,"Q110740"
,"Q109367","Q109244"
,"Q108950","Q108855","Q108617","Q108856","Q108754","Q108342","Q108343"
,"Q107869","Q107491"
,"Q106993","Q106997","Q106272","Q106388","Q106389","Q106042"
,"Q105840","Q105655"
,"Q104996"
,"Q103293"
,"Q102906","Q102674","Q102687","Q102289","Q102089"
,"Q101162","Q101163","Q101596"
,"Q100689","Q100680","Q100562","Q100010"
,"Q99982"
,"Q99716"
,"Q99581"
,"Q99480"
,"Q98869"
,"Q98578"
,"Q98197"
,"Q98059","Q98078"
,"Q96024" # Done
,".pos")
if (glb_rsp_var_raw != glb_rsp_var)
glbFeatsExclude <- union(glbFeatsExclude, glb_rsp_var_raw)
glbFeatsInteractionOnly <- list()
#glbFeatsInteractionOnly[["<child_feat>"]] <- "<parent_feat>"
glbFeatsInteractionOnly[["YOB.Age.dff"]] <- "YOB.Age.fctr"
glbFeatsDrop <- c(NULL
# , "<feat1>", "<feat2>"
)
glb_map_vars <- NULL # or c("<var1>", "<var2>")
glb_map_urls <- list();
# glb_map_urls[["<var1>"]] <- "<var1.url>"
# Derived features; Use this mechanism to cleanse data ??? Cons: Data duplication ???
glbFeatsDerive <- list();
# glbFeatsDerive[["<feat.my.sfx>"]] <- list(
# mapfn = function(<arg1>, <arg2>) { return(function(<arg1>, <arg2>)) }
# , args = c("<arg1>", "<arg2>"))
#myprint_df(data.frame(ImageId = mapfn(glbObsAll$.src, glbObsAll$.pos)))
#data.frame(ImageId = mapfn(glbObsAll$.src, glbObsAll$.pos))[7045:7055, ]
# character
# mapfn = function(Education) { raw <- Education; raw[is.na(raw)] <- "NA.my"; return(as.factor(raw)) }
# mapfn = function(Week) { return(substr(Week, 1, 10)) }
# mapfn = function(Name) { return(sapply(Name, function(thsName)
# str_sub(unlist(str_split(thsName, ","))[1], 1, 1))) }
# mapfn = function(descriptor) { return(plyr::revalue(descriptor, c(
# "ABANDONED BUILDING" = "OTHER",
# "**" = "**"
# ))) }
# mapfn = function(description) { mod_raw <- description;
# This is here because it does not work if it's in txt_map_filename
# mod_raw <- gsub(paste0(c("\n", "\211", "\235", "\317", "\333"), collapse = "|"), " ", mod_raw)
# Don't parse for "." because of ".com"; use customized gsub for that text
# mod_raw <- gsub("(\\w)(!|\\*|,|-|/)(\\w)", "\\1\\2 \\3", mod_raw);
# Some state acrnoyms need context for separation e.g.
# LA/L.A. could either be "Louisiana" or "LosAngeles"
# modRaw <- gsub("\\bL\\.A\\.( |,|')", "LosAngeles\\1", modRaw);
# OK/O.K. could either be "Oklahoma" or "Okay"
# modRaw <- gsub("\\bACA OK\\b", "ACA OKay", modRaw);
# modRaw <- gsub("\\bNow O\\.K\\.\\b", "Now OKay", modRaw);
# PR/P.R. could either be "PuertoRico" or "Public Relations"
# modRaw <- gsub("\\bP\\.R\\. Campaign", "PublicRelations Campaign", modRaw);
# VA/V.A. could either be "Virginia" or "VeteransAdministration"
# modRaw <- gsub("\\bthe V\\.A\\.\\:", "the VeteranAffairs:", modRaw);
#
# Custom mods
# return(mod_raw) }
# numeric
# Create feature based on record position/id in data
glbFeatsDerive[[".pos"]] <- list(
mapfn = function(raw1) { return(1:length(raw1)) }
, args = c(".rnorm"))
# glbFeatsDerive[[".pos.y"]] <- list(
# mapfn = function(raw1) { return(1:length(raw1)) }
# , args = c(".rnorm"))
# Add logs of numerics that are not distributed normally
# Derive & keep multiple transformations of the same feature, if normality is hard to achieve with just one transformation
# Right skew: logp1; sqrt; ^ 1/3; logp1(logp1); log10; exp(-<feat>/constant)
# glbFeatsDerive[["WordCount.log1p"]] <- list(
# mapfn = function(WordCount) { return(log1p(WordCount)) }
# , args = c("WordCount"))
# glbFeatsDerive[["WordCount.root2"]] <- list(
# mapfn = function(WordCount) { return(WordCount ^ (1/2)) }
# , args = c("WordCount"))
# glbFeatsDerive[["WordCount.nexp"]] <- list(
# mapfn = function(WordCount) { return(exp(-WordCount)) }
# , args = c("WordCount"))
#print(summary(glbObsAll$WordCount))
#print(summary(mapfn(glbObsAll$WordCount)))
# If imputation shd be skipped for this feature
# glbFeatsDerive[["District.fctr"]] <- list(
# mapfn = function(District) {
# raw <- District;
# ret_vals <- rep_len("NA", length(raw));
# ret_vals[!is.na(raw)] <- sapply(raw[!is.na(raw)], function(elm)
# ifelse(elm < 10, "1-9",
# ifelse(elm < 20, "10-19", "20+")));
# return(relevel(as.factor(ret_vals), ref = "NA"))
# }
# , args = c("District"))
# YOB options:
# 1. Missing data:
# 1.1 0 -> Does not improve baseline
# 1.2 Cut factors & "NA" is a level
# 2. Data corrections: < 1928 & > 2000
# 3. Scale YOB
# 4. Add Age
# YOB.Age.fctr needs to be synced with YOB.Age.dff; Create a separate sub-function ???
glbFeatsDerive[["YOB.Age.fctr"]] <- list(
mapfn = function(raw1) {
raw <- 2016 - raw1
# raw[!is.na(raw) & raw >= 2010] <- NA
raw[!is.na(raw) & (raw <= 15)] <- NA
raw[!is.na(raw) & (raw >= 90)] <- NA
retVal <- rep_len("NA", length(raw))
# breaks = c(1879, seq(1949, 1989, 10), 2049)
# cutVal <- cut(raw[!is.na(raw)], breaks = breaks,
# labels = as.character(breaks + 1)[1:(length(breaks) - 1)])
cutVal <- cut(raw[!is.na(raw)], breaks = c(15, 20, 25, 30, 35, 40, 50, 65, 90))
retVal[!is.na(raw)] <- levels(cutVal)[cutVal]
return(factor(retVal, levels = c("NA"
,"(15,20]","(20,25]","(25,30]","(30,35]","(35,40]","(40,50]","(50,65]","(65,90]"),
ordered = TRUE))
}
, args = c("YOB"))
# YOB.Age.fctr needs to be synced with YOB.Age.dff; Create a separate sub-function ???
glbFeatsDerive[["YOB.Age.dff"]] <- list(
mapfn = function(raw1) {
raw <- 2016 - raw1
raw[!is.na(raw) & (raw <= 15)] <- NA
raw[!is.na(raw) & (raw >= 90)] <- NA
breaks <- c(15, 20, 25, 30, 35, 40, 50, 65, 90)
# retVal <- rep_len(0, length(raw))
stopifnot(sum(!is.na(raw) && (raw <= 15)) == 0)
stopifnot(sum(!is.na(raw) && (raw >= 90)) == 0)
# msk <- !is.na(raw) && (raw > 15) && (raw <= 20); if (sum(msk > 0)) retVal[msk] <- raw[msk] - 15
# msk <- !is.na(raw) && (raw > 20) && (raw <= 25); if (sum(msk > 0)) retVal[msk] <- raw[msk] - 20
# msk <- !is.na(raw) && (raw > 25) && (raw <= 30); if (sum(msk > 0)) retVal[msk] <- raw[msk] - 25
# msk <- !is.na(raw) && (raw > 30) && (raw <= 35); if (sum(msk > 0)) retVal[msk] <- raw[msk] - 30
# msk <- !is.na(raw) && (raw > 35) && (raw <= 40); if (sum(msk > 0)) retVal[msk] <- raw[msk] - 35
# msk <- !is.na(raw) && (raw > 40) && (raw <= 50); if (sum(msk > 0)) retVal[msk] <- raw[msk] - 40
# msk <- !is.na(raw) && (raw > 50) && (raw <= 65); if (sum(msk > 0)) retVal[msk] <- raw[msk] - 50
# msk <- !is.na(raw) && (raw > 65) && (raw <= 90); if (sum(msk > 0)) retVal[msk] <- raw[msk] - 65
breaks <- c(15, 20, 25, 30, 35, 40, 50, 65, 90)
retVal <- sapply(raw, function(age) {
if (is.na(age)) return(0) else
if ((age > 15) && (age <= 20)) return(age - 15) else
if ((age > 20) && (age <= 25)) return(age - 20) else
if ((age > 25) && (age <= 30)) return(age - 25) else
if ((age > 30) && (age <= 35)) return(age - 30) else
if ((age > 35) && (age <= 40)) return(age - 35) else
if ((age > 40) && (age <= 50)) return(age - 40) else
if ((age > 50) && (age <= 65)) return(age - 50) else
if ((age > 65) && (age <= 90)) return(age - 65)
})
return(retVal)
}
, args = c("YOB"))
glbFeatsDerive[["Gender.fctr"]] <- list(
mapfn = function(raw1) {
raw <- raw1
raw[raw %in% ""] <- "N"
raw <- gsub("Male" , "M", raw, fixed = TRUE)
raw <- gsub("Female", "F", raw, fixed = TRUE)
return(relevel(as.factor(raw), ref = "N"))
}
, args = c("Gender"))
glbFeatsDerive[["Income.fctr"]] <- list(
mapfn = function(raw1) { raw <- raw1;
raw[raw %in% ""] <- "N"
raw <- gsub("under $25,000" , "<25K" , raw, fixed = TRUE)
raw <- gsub("$25,001 - $50,000" , "25-50K" , raw, fixed = TRUE)
raw <- gsub("$50,000 - $74,999" , "50-75K" , raw, fixed = TRUE)
raw <- gsub("$75,000 - $100,000" , "75-100K" , raw, fixed = TRUE)
raw <- gsub("$100,001 - $150,000", "100-150K", raw, fixed = TRUE)
raw <- gsub("over $150,000" , ">150K" , raw, fixed = TRUE)
return(factor(raw, levels = c("N","<25K","25-50K","50-75K","75-100K","100-150K",">150K"),
ordered = TRUE))
}
, args = c("Income"))
glbFeatsDerive[["Hhold.fctr"]] <- list(
mapfn = function(raw1) { raw <- raw1;
raw[raw %in% ""] <- "N"
raw <- gsub("Domestic Partners (no kids)", "PKn", raw, fixed = TRUE)
raw <- gsub("Domestic Partners (w/kids)" , "PKy", raw, fixed = TRUE)
raw <- gsub("Married (no kids)" , "MKn", raw, fixed = TRUE)
raw <- gsub("Married (w/kids)" , "MKy", raw, fixed = TRUE)
raw <- gsub("Single (no kids)" , "SKn", raw, fixed = TRUE)
raw <- gsub("Single (w/kids)" , "SKy", raw, fixed = TRUE)
return(relevel(as.factor(raw), ref = "N"))
}
, args = c("HouseholdStatus"))
glbFeatsDerive[["Edn.fctr"]] <- list(
mapfn = function(raw1) { raw <- raw1;
raw[raw %in% ""] <- "N"
raw <- gsub("Current K-12" , "K12", raw, fixed = TRUE)
raw <- gsub("High School Diploma" , "HSD", raw, fixed = TRUE)
raw <- gsub("Current Undergraduate", "CCg", raw, fixed = TRUE)
raw <- gsub("Associate's Degree" , "Ast", raw, fixed = TRUE)
raw <- gsub("Bachelor's Degree" , "Bcr", raw, fixed = TRUE)
raw <- gsub("Master's Degree" , "Msr", raw, fixed = TRUE)
raw <- gsub("Doctoral Degree" , "PhD", raw, fixed = TRUE)
return(factor(raw, levels = c("N","K12","HSD","CCg","Ast","Bcr","Msr","PhD"),
ordered = TRUE))
}
, args = c("EducationLevel"))
# for (qsn in c("Q124742","Q124122"))
# for (qsn in grep("Q12(.{4})(?!\\.fctr)", names(glbObsTrn), value = TRUE, perl = TRUE))
for (qsn in grep("Q", glbFeatsExclude, fixed = TRUE, value = TRUE))
glbFeatsDerive[[paste0(qsn, ".fctr")]] <- list(
mapfn = function(raw1) {
raw1[raw1 %in% ""] <- "NA"
rawVal <- unique(raw1)
if (length(setdiff(rawVal, (expVal <- c("NA", "No", "Ys")))) == 0) {
raw1 <- gsub("Yes", "Ys", raw1, fixed = TRUE)
if (length(setdiff(rawVal, expVal)) > 0)
stop(qsn, " vals: ", paste0(rawVal, collapse = "|"),
" does not match expectation: ", paste0(expVal, collapse = "|"))
} else
if (length(setdiff(rawVal, (expVal <- c("NA", "Me", "Circumstances")))) == 0) {
raw1 <- gsub("Circumstances", "Cs", raw1, fixed = TRUE)
if (length(setdiff(rawVal, expVal)) > 0)
stop(qsn, " vals: ", paste0(rawVal, collapse = "|"),
" does not match expectation: ", paste0(expVal, collapse = "|"))
} else
if (length(setdiff(rawVal, (expVal <- c("NA", "Grrr people", "Yay people!")))) == 0) {
raw1 <- gsub("Grrr people", "Gr", raw1, fixed = TRUE)
raw1 <- gsub("Yay people!", "Yy", raw1, fixed = TRUE)
if (length(setdiff(rawVal, expVal)) > 0)
stop(qsn, " vals: ", paste0(rawVal, collapse = "|"),
" does not match expectation: ", paste0(expVal, collapse = "|"))
} else
if (length(setdiff(rawVal, (expVal <- c("NA", "Idealist", "Pragmatist")))) == 0) {
raw1 <- gsub("Idealist" , "Id", raw1, fixed = TRUE)
raw1 <- gsub("Pragmatist", "Pr", raw1, fixed = TRUE)
if (length(setdiff(rawVal, expVal)) > 0)
stop(qsn, " vals: ", paste0(rawVal, collapse = "|"),
" does not match expectation: ", paste0(expVal, collapse = "|"))
} else
if (length(setdiff(rawVal, (expVal <- c("NA", "Private", "Public")))) == 0) {
raw1 <- gsub("Private", "Pt", raw1, fixed = TRUE)
raw1 <- gsub("Public" , "Pc", raw1, fixed = TRUE)
if (length(setdiff(rawVal, expVal)) > 0)
stop(qsn, " vals: ", paste0(rawVal, collapse = "|"),
" does not match expectation: ", paste0(expVal, collapse = "|"))
}
return(relevel(as.factor(raw1), ref = "NA"))
}
, args = c(qsn))
# If imputation of missing data is not working ...
# glbFeatsDerive[["FertilityRate.nonNA"]] <- list(
# mapfn = function(FertilityRate, Region) {
# RegionMdn <- tapply(FertilityRate, Region, FUN = median, na.rm = TRUE)
#
# retVal <- FertilityRate
# retVal[is.na(FertilityRate)] <- RegionMdn[Region[is.na(FertilityRate)]]
# return(retVal)
# }
# , args = c("FertilityRate", "Region"))
# mapfn = function(HOSPI.COST) { return(cut(HOSPI.COST, 5, breaks = c(0, 100000, 200000, 300000, 900000), labels = NULL)) }
# mapfn = function(Rasmussen) { return(ifelse(sign(Rasmussen) >= 0, 1, 0)) }
# mapfn = function(startprice) { return(startprice ^ (1/2)) }
# mapfn = function(startprice) { return(log(startprice)) }
# mapfn = function(startprice) { return(exp(-startprice / 20)) }
# mapfn = function(startprice) { return(scale(log(startprice))) }
# mapfn = function(startprice) { return(sign(sprice.predict.diff) * (abs(sprice.predict.diff) ^ (1/10))) }
# factor
# mapfn = function(PropR) { return(as.factor(ifelse(PropR >= 0.5, "Y", "N"))) }
# mapfn = function(productline, description) { as.factor(gsub(" ", "", productline)) }
# mapfn = function(purpose) { return(relevel(as.factor(purpose), ref="all_other")) }
# mapfn = function(raw) { tfr_raw <- as.character(cut(raw, 5));
# tfr_raw[is.na(tfr_raw)] <- "NA.my";
# return(as.factor(tfr_raw)) }
# mapfn = function(startprice.log10) { return(cut(startprice.log10, 3)) }
# mapfn = function(startprice.log10) { return(cut(sprice.predict.diff, c(-1000, -100, -10, -1, 0, 1, 10, 100, 1000))) }
# , args = c("<arg1>"))
# multiple args
# mapfn = function(id, date) { return(paste(as.character(id), as.character(date), sep = "#")) }
# mapfn = function(PTS, oppPTS) { return(PTS - oppPTS) }
# mapfn = function(startprice.log10.predict, startprice) {
# return(spdiff <- (10 ^ startprice.log10.predict) - startprice) }
# mapfn = function(productline, description) { as.factor(
# paste(gsub(" ", "", productline), as.numeric(nchar(description) > 0), sep = "*")) }
# mapfn = function(.src, .pos) {
# return(paste(.src, sprintf("%04d",
# ifelse(.src == "Train", .pos, .pos - 7049)
# ), sep = "#")) }
# # If glbObsAll is not sorted in the desired manner
# mapfn=function(Week) { return(coredata(lag(zoo(orderBy(~Week, glbObsAll)$ILI), -2, na.pad=TRUE))) }
# mapfn=function(ILI) { return(coredata(lag(zoo(ILI), -2, na.pad=TRUE))) }
# mapfn=function(ILI.2.lag) { return(log(ILI.2.lag)) }
# glbFeatsDerive[["<var1>"]] <- glbFeatsDerive[["<var2>"]]
# tst <- "descr.my"; args_lst <- NULL; for (arg in glbFeatsDerive[[tst]]$args) args_lst[[arg]] <- glbObsAll[, arg]; print(head(args_lst[[arg]])); print(head(drv_vals <- do.call(glbFeatsDerive[[tst]]$mapfn, args_lst)));
# print(which_ix <- which(args_lst[[arg]] == 0.75)); print(drv_vals[which_ix]);
glbFeatsDateTime <- list()
# Use OlsonNames() to enumerate supported time zones
# glbFeatsDateTime[["<DateTimeFeat>"]] <-
# c(format = "%Y-%m-%d %H:%M:%S" or "%m/%e/%y", timezone = "US/Eastern", impute.na = TRUE,
# last.ctg = FALSE, poly.ctg = FALSE)
glbFeatsPrice <- NULL # or c("<price_var>")
glbFeatsImage <- list() #list(<imageFeat> = list(patchSize = 10)) # if patchSize not specified, no patch computation
glbFeatsText <- list()
Sys.setlocale("LC_ALL", "C") # For english## [1] "C/C/C/C/C/en_US.UTF-8"#glbFeatsText[["<TextFeature>"]] <- list(NULL,
# ,names = myreplacePunctuation(str_to_lower(gsub(" ", "", c(NULL,
# <comma-separated-screened-names>
# ))))
# ,rareWords = myreplacePunctuation(str_to_lower(gsub(" ", "", c(NULL,
# <comma-separated-nonSCOWL-words>
# ))))
#)
# Text Processing Step: custom modifications not present in txt_munge -> use glbFeatsDerive
# Text Processing Step: universal modifications
glb_txt_munge_filenames_pfx <- "<projectId>_mytxt_"
# Text Processing Step: tolower
# Text Processing Step: myreplacePunctuation
# Text Processing Step: removeWords
glb_txt_stop_words <- list()
# Remember to use unstemmed words
if (length(glbFeatsText) > 0) {
require(tm)
require(stringr)
glb_txt_stop_words[["<txt_var>"]] <- sort(myreplacePunctuation(str_to_lower(gsub(" ", "", c(NULL
# Remove any words from stopwords
# , setdiff(myreplacePunctuation(stopwords("english")), c("<keep_wrd1>", <keep_wrd2>"))
# Remove salutations
,"mr","mrs","dr","Rev"
# Remove misc
#,"th" # Happy [[:digit::]]+th birthday
# Remove terms present in Trn only or New only; search for "Partition post-stem"
# ,<comma-separated-terms>
# cor.y.train == NA
# ,unlist(strsplit(paste(c(NULL
# ,"<comma-separated-terms>"
# ), collapse=",")
# freq == 1; keep c("<comma-separated-terms-to-keep>")
# ,<comma-separated-terms>
# chisq.pval high (e.g. == 1); keep c("<comma-separated-terms-to-keep>")
# ,<comma-separated-terms>
# nzv.freqRatio high (e.g. >= glbFeatsNzvFreqMax); keep c("<comma-separated-terms-to-keep>")
# ,<comma-separated-terms>
)))))
}
#orderBy(~term, glb_post_stem_words_terms_df_lst[[txtFeat]][grep("^man", glb_post_stem_words_terms_df_lst[[txtFeat]]$term), ])
#glbObsAll[glb_post_stem_words_terms_mtrx_lst[[txtFeat]][, 4866] > 0, c(glb_rsp_var, txtFeat)]
# To identify terms with a specific freq
#paste0(sort(subset(glb_post_stop_words_terms_df_lst[[txtFeat]], freq == 1)$term), collapse = ",")
#paste0(sort(subset(glb_post_stem_words_terms_df_lst[[txtFeat]], freq <= 2)$term), collapse = ",")
#subset(glb_post_stem_words_terms_df_lst[[txtFeat]], term %in% c("zinger"))
# To identify terms with a specific freq &
# are not stemmed together later OR is value of color.fctr (e.g. gold)
#paste0(sort(subset(glb_post_stop_words_terms_df_lst[[txtFeat]], (freq == 1) & !(term %in% c("blacked","blemish","blocked","blocks","buying","cables","careful","carefully","changed","changing","chargers","cleanly","cleared","connect","connects","connected","contains","cosmetics","default","defaulting","defective","definitely","describe","described","devices","displays","drop","drops","engravement","excellant","excellently","feels","fix","flawlessly","frame","framing","gentle","gold","guarantee","guarantees","handled","handling","having","install","iphone","iphones","keeped","keeps","known","lights","line","lining","liquid","liquidation","looking","lots","manuals","manufacture","minis","most","mostly","network","networks","noted","opening","operated","performance","performs","person","personalized","photograph","physically","placed","places","powering","pre","previously","products","protection","purchasing","returned","rotate","rotation","running","sales","second","seconds","shipped","shuts","sides","skin","skinned","sticker","storing","thats","theres","touching","unusable","update","updates","upgrade","weeks","wrapped","verified","verify") ))$term), collapse = ",")
#print(subset(glb_post_stem_words_terms_df_lst[[txtFeat]], (freq <= 2)))
#glbObsAll[which(terms_mtrx[, 229] > 0), glbFeatsText]
# To identify terms with cor.y == NA
#orderBy(~-freq+term, subset(glb_post_stop_words_terms_df_lst[[txtFeat]], is.na(cor.y)))
#paste(sort(subset(glb_post_stop_words_terms_df_lst[[txtFeat]], is.na(cor.y))[, "term"]), collapse=",")
#orderBy(~-freq+term, subset(glb_post_stem_words_terms_df_lst[[txtFeat]], is.na(cor.y)))
# To identify terms with low cor.y.abs
#head(orderBy(~cor.y.abs+freq+term, subset(glb_post_stem_words_terms_df_lst[[txtFeat]], !is.na(cor.y))), 5)
# To identify terms with high chisq.pval
#subset(glb_post_stem_words_terms_df_lst[[txtFeat]], chisq.pval > 0.99)
#paste0(sort(subset(glb_post_stem_words_terms_df_lst[[txtFeat]], (chisq.pval > 0.99) & (freq <= 10))$term), collapse=",")
#paste0(sort(subset(glb_post_stem_words_terms_df_lst[[txtFeat]], (chisq.pval > 0.9))$term), collapse=",")
#head(orderBy(~-chisq.pval+freq+term, glb_post_stem_words_terms_df_lst[[txtFeat]]), 5)
#glbObsAll[glb_post_stem_words_terms_mtrx_lst[[txtFeat]][, 68] > 0, glbFeatsText]
#orderBy(~term, glb_post_stem_words_terms_df_lst[[txtFeat]][grep("^m", glb_post_stem_words_terms_df_lst[[txtFeat]]$term), ])
# To identify terms with high nzv.freqRatio
#summary(glb_post_stem_words_terms_df_lst[[txtFeat]]$nzv.freqRatio)
#paste0(sort(setdiff(subset(glb_post_stem_words_terms_df_lst[[txtFeat]], (nzv.freqRatio >= glbFeatsNzvFreqMax) & (freq < 10) & (chisq.pval >= 0.05))$term, c( "128gb","3g","4g","gold","ipad1","ipad3","ipad4","ipadair2","ipadmini2","manufactur","spacegray","sprint","tmobil","verizon","wifion"))), collapse=",")
# To identify obs with a txt term
#tail(orderBy(~-freq+term, glb_post_stop_words_terms_df_lst[[txtFeat]]), 20)
#mydspObs(list(descr.my.contains="non"), cols=c("color", "carrier", "cellular", "storage"))
#grep("ever", dimnames(terms_stop_mtrx)$Terms)
#which(terms_stop_mtrx[, grep("ipad", dimnames(terms_stop_mtrx)$Terms)] > 0)
#glbObsAll[which(terms_stop_mtrx[, grep("16", dimnames(terms_stop_mtrx)$Terms)[1]] > 0), c(glbFeatsCategory, "storage", txtFeat)]
# Text Processing Step: screen for names # Move to glbFeatsText specs section in order of text processing steps
# glbFeatsText[["<txtFeat>"]]$names <- myreplacePunctuation(str_to_lower(gsub(" ", "", c(NULL
# # Person names for names screening
# ,<comma-separated-list>
#
# # Company names
# ,<comma-separated-list>
#
# # Product names
# ,<comma-separated-list>
# ))))
# glbFeatsText[["<txtFeat>"]]$rareWords <- myreplacePunctuation(str_to_lower(gsub(" ", "", c(NULL
# # Words not in SCOWL db
# ,<comma-separated-list>
# ))))
# To identify char vectors post glbFeatsTextMap
#grep("six(.*)hour", glb_txt_chr_lst[[txtFeat]], ignore.case = TRUE, value = TRUE)
#grep("[S|s]ix(.*)[H|h]our", glb_txt_chr_lst[[txtFeat]], value = TRUE)
# To identify whether terms shd be synonyms
#orderBy(~term, glb_post_stop_words_terms_df_lst[[txtFeat]][grep("^moder", glb_post_stop_words_terms_df_lst[[txtFeat]]$term), ])
# term_row_df <- glb_post_stop_words_terms_df_lst[[txtFeat]][grep("^came$", glb_post_stop_words_terms_df_lst[[txtFeat]]$term), ]
#
# cor(glb_post_stop_words_terms_mtrx_lst[[txtFeat]][glbObsAll$.lcn == "Fit", term_row_df$pos], glbObsTrn[, glb_rsp_var], use="pairwise.complete.obs")
# To identify which stopped words are "close" to a txt term
#sort(glbFeatsCluster)
# Text Processing Step: stemDocument
# To identify stemmed txt terms
#glb_post_stop_words_terms_df_lst[[txtFeat]][grep("^la$", glb_post_stop_words_terms_df_lst[[txtFeat]]$term), ]
#orderBy(~term, glb_post_stem_words_terms_df_lst[[txtFeat]][grep("^con", glb_post_stem_words_terms_df_lst[[txtFeat]]$term), ])
#glbObsAll[which(terms_stem_mtrx[, grep("use", dimnames(terms_stem_mtrx)$Terms)[[1]]] > 0), c(glbFeatsId, "productline", txtFeat)]
#glbObsAll[which(TfIdf_stem_mtrx[, 191] > 0), c(glbFeatsId, glbFeatsCategory, txtFeat)]
#glbObsAll[which(glb_post_stop_words_terms_mtrx_lst[[txtFeat]][, 6165] > 0), c(glbFeatsId, glbFeatsCategory, txtFeat)]
#which(glbObsAll$UniqueID %in% c(11915, 11926, 12198))
# Text Processing Step: mycombineSynonyms
# To identify which terms are associated with not -> combine "could not" & "couldn't"
#findAssocs(glb_full_DTM_lst[[txtFeat]], "not", 0.05)
# To identify which synonyms should be combined
#orderBy(~term, glb_post_stem_words_terms_df_lst[[txtFeat]][grep("^c", glb_post_stem_words_terms_df_lst[[txtFeat]]$term), ])
chk_comb_cor <- function(syn_lst) {
# cor(terms_stem_mtrx[glbObsAll$.src == "Train", grep("^(damag|dent|ding)$", dimnames(terms_stem_mtrx)[[2]])], glbObsTrn[, glb_rsp_var], use="pairwise.complete.obs")
print(subset(glb_post_stem_words_terms_df_lst[[txtFeat]], term %in% syn_lst$syns))
print(subset(get_corpus_terms(tm_map(glbFeatsTextCorpus[[txtFeat]], mycombineSynonyms, list(syn_lst), lazy=FALSE)), term == syn_lst$word))
# cor(terms_stop_mtrx[glbObsAll$.src == "Train", grep("^(damage|dent|ding)$", dimnames(terms_stop_mtrx)[[2]])], glbObsTrn[, glb_rsp_var], use="pairwise.complete.obs")
# cor(rowSums(terms_stop_mtrx[glbObsAll$.src == "Train", grep("^(damage|dent|ding)$", dimnames(terms_stop_mtrx)[[2]])]), glbObsTrn[, glb_rsp_var], use="pairwise.complete.obs")
}
#chk_comb_cor(syn_lst=list(word="cabl", syns=c("cabl", "cord")))
#chk_comb_cor(syn_lst=list(word="damag", syns=c("damag", "dent", "ding")))
#chk_comb_cor(syn_lst=list(word="dent", syns=c("dent", "ding")))
#chk_comb_cor(syn_lst=list(word="use", syns=c("use", "usag")))
glbFeatsTextSynonyms <- list()
# list parsed to collect glbFeatsText[[<txtFeat>]]$vldTerms
# glbFeatsTextSynonyms[["Hdln.my"]] <- list(NULL
# # people in places
# , list(word = "australia", syns = c("australia", "australian"))
# , list(word = "italy", syns = c("italy", "Italian"))
# , list(word = "newyork", syns = c("newyork", "newyorker"))
# , list(word = "Pakistan", syns = c("Pakistan", "Pakistani"))
# , list(word = "peru", syns = c("peru", "peruvian"))
# , list(word = "qatar", syns = c("qatar", "qatari"))
# , list(word = "scotland", syns = c("scotland", "scotish"))
# , list(word = "Shanghai", syns = c("Shanghai", "Shanzhai"))
# , list(word = "venezuela", syns = c("venezuela", "venezuelan"))
#
# # companies - needs to be data dependent
# # - e.g. ensure BNP in this experiment/feat always refers to BNPParibas
#
# # general synonyms
# , list(word = "Create", syns = c("Create","Creator"))
# , list(word = "cute", syns = c("cute","cutest"))
# , list(word = "Disappear", syns = c("Disappear","Fadeout"))
# , list(word = "teach", syns = c("teach", "taught"))
# , list(word = "theater", syns = c("theater", "theatre", "theatres"))
# , list(word = "understand", syns = c("understand", "understood"))
# , list(word = "weak", syns = c("weak", "weaken", "weaker", "weakest"))
# , list(word = "wealth", syns = c("wealth", "wealthi"))
#
# # custom synonyms (phrases)
#
# # custom synonyms (names)
# )
#glbFeatsTextSynonyms[["<txtFeat>"]] <- list(NULL
# , list(word="<stem1>", syns=c("<stem1>", "<stem1_2>"))
# )
for (txtFeat in names(glbFeatsTextSynonyms))
for (entryIx in 1:length(glbFeatsTextSynonyms[[txtFeat]])) {
glbFeatsTextSynonyms[[txtFeat]][[entryIx]]$word <-
str_to_lower(glbFeatsTextSynonyms[[txtFeat]][[entryIx]]$word)
glbFeatsTextSynonyms[[txtFeat]][[entryIx]]$syns <-
str_to_lower(glbFeatsTextSynonyms[[txtFeat]][[entryIx]]$syns)
}
glbFeatsTextSeed <- 181
# tm options include: check tm::weightSMART
glb_txt_terms_control <- list( # Gather model performance & run-time stats
# weighting = function(x) weightSMART(x, spec = "nnn")
# weighting = function(x) weightSMART(x, spec = "lnn")
# weighting = function(x) weightSMART(x, spec = "ann")
# weighting = function(x) weightSMART(x, spec = "bnn")
# weighting = function(x) weightSMART(x, spec = "Lnn")
#
weighting = function(x) weightSMART(x, spec = "ltn") # default
# weighting = function(x) weightSMART(x, spec = "lpn")
#
# weighting = function(x) weightSMART(x, spec = "ltc")
#
# weighting = weightBin
# weighting = weightTf
# weighting = weightTfIdf # : default
# termFreq selection criteria across obs: tm default: list(global=c(1, Inf))
, bounds = list(global = c(1, Inf))
# wordLengths selection criteria: tm default: c(3, Inf)
, wordLengths = c(1, Inf)
)
glb_txt_cor_var <- glb_rsp_var # : default # or c(<feat>)
# select one from c("union.top.val.cor", "top.cor", "top.val", default: "top.chisq", "sparse")
glbFeatsTextFilter <- "top.chisq"
glbFeatsTextTermsMax <- rep(10, length(glbFeatsText)) # :default
names(glbFeatsTextTermsMax) <- names(glbFeatsText)
# Text Processing Step: extractAssoc
glbFeatsTextAssocCor <- rep(1, length(glbFeatsText)) # :default
names(glbFeatsTextAssocCor) <- names(glbFeatsText)
# Remember to use stemmed terms
glb_important_terms <- list()
# Text Processing Step: extractPatterns (ngrams)
glbFeatsTextPatterns <- list()
#glbFeatsTextPatterns[[<txtFeat>>]] <- list()
#glbFeatsTextPatterns[[<txtFeat>>]] <- c(metropolitan.diary.colon = "Metropolitan Diary:")
# Have to set it even if it is not used
# Properties:
# numrows(glb_feats_df) << numrows(glbObsFit
# Select terms that appear in at least 0.2 * O(FP/FN(glbObsOOB)) ???
# numrows(glbObsOOB) = 1.1 * numrows(glbObsNew) ???
glb_sprs_thresholds <- NULL # or c(<txtFeat1> = 0.988, <txtFeat2> = 0.970, <txtFeat3> = 0.970)
glbFctrMaxUniqVals <- 20 # default: 20
glb_impute_na_data <- FALSE # or TRUE
glb_mice_complete.seed <- 144 # or any integer
glbFeatsCluster <- paste(grep("^Q.", glbFeatsExclude, value = TRUE), "fctr", sep = ".") # NULL : glbFeatsCluster <- c("YOB.Age.fctr", "Gender.fctr", "Income.fctr",
# # "Hhold.fctr",
# "Edn.fctr",
# paste(grep("^Q.", glbFeatsExclude, value = TRUE), "fctr", sep = ".")) # NULL : default or c("<feat1>", "<feat2>")
# glbFeatsCluster <- grep(paste0("[",
# toupper(paste0(substr(glbFeatsText, 1, 1), collapse = "")),
# "]\\.[PT]\\."),
# names(glbObsAll), value = TRUE)
glb_cluster.seed <- 189 # or any integer
glbClusterEntropyVar <- NULL # c(glb_rsp_var, as.factor(cut(glb_rsp_var, 3)), default: NULL)
glbFeatsClusterVarsExclude <- FALSE # default FALSE
glb_interaction_only_feats <- NULL # : default or c(<parent_feat> = "<child_feat>")
glbFeatsNzvFreqMax <- 19 # 19 : caret default
glbFeatsNzvUniqMin <- 10 # 10 : caret default
glbRFESizes <- list()
#glbRFESizes[["mdlFamily"]] <- c(4, 8, 16, 32, 64, 67, 68, 69) # Accuracy@69/70 = 0.8258
glbRFESizes[["RFE.X"]] <- c(96, 112, 120, 124, 128, 129, 130, 131, 132, 133, 135, 138, 142, 157, 187, 247) # accuracy(131) = 0.6285
glbRFESizes[["Final"]] <- c(8, 16, 32, 40, 44, 46, 48, 49, 50, 51, 52, 56, 64, 96, 128, 247) # accuracy(49) = 0.6164
glbRFEResults <- NULL
glbObsFitOutliers <- list()
# If outliers.n >= 10; consider concatenation of interaction vars
# glbObsFitOutliers[["<mdlFamily>"]] <- c(NULL
# is.na(.rstudent)
# max(.rstudent)
# is.na(.dffits)
# .hatvalues >= 0.99
# -38,167,642 < minmax(.rstudent) < 49,649,823
# , <comma-separated-<glbFeatsId>>
# )
glbObsTrnOutliers <- list()
glbObsTrnOutliers[["Final"]] <- union(glbObsFitOutliers[["All.X"]],
c(NULL
))
# Modify mdlId to (build & extract) "<FamilyId>#<Fit|Trn>#<caretMethod>#<preProc1.preProc2>#<samplingMethod>"
glb_models_lst <- list(); glb_models_df <- data.frame()
# Add xgboost algorithm
# Regression
if (glb_is_regression) {
glbMdlMethods <- c(NULL
# deterministic
#, "lm", # same as glm
, "glm", "bayesglm", "glmnet"
, "rpart"
# non-deterministic
, "gbm", "rf"
# Unknown
, "nnet" , "avNNet" # runs 25 models per cv sample for tunelength=5
, "svmLinear", "svmLinear2"
, "svmPoly" # runs 75 models per cv sample for tunelength=5
, "svmRadial"
, "earth"
, "bagEarth" # Takes a long time
,"xgbLinear","xgbTree"
)
} else
# Classification - Add ada (auto feature selection)
if (glb_is_binomial)
glbMdlMethods <- c(NULL
# deterministic
, "bagEarth" # Takes a long time
, "glm", "bayesglm", "glmnet"
, "nnet"
, "rpart"
# non-deterministic
, "gbm"
, "avNNet" # runs 25 models per cv sample for tunelength=5
, "rf"
# Unknown
, "lda", "lda2"
# svm models crash when predict is called -> internal to kernlab it should call predict without .outcome
, "svmLinear", "svmLinear2"
, "svmPoly" # runs 75 models per cv sample for tunelength=5
, "svmRadial"
, "earth"
,"xgbLinear","xgbTree"
) else
glbMdlMethods <- c(NULL
# deterministic
,"glmnet"
# non-deterministic
,"rf"
# Unknown
,"gbm","rpart","xgbLinear","xgbTree"
)
glbMdlFamilies <- list(); glb_mdl_feats_lst <- list()
# family: Choose from c("RFE.X", "Csm.X", "All.X", "Best.Interact") %*% c(NUll, ".NOr", "Inc")
# RFE = "Recursive Feature Elimination"
# Csm = CuStoM
# NOr = No OutlieRs
# Inc = INteraCt
# methods: Choose from c(NULL, <method>, glbMdlMethods)
#glbMdlFamilies[["RFE.X"]] <- c("glmnet", "glm") # non-NULL vector is mandatory
if (glb_is_classification && !glb_is_binomial) {
# glm does not work for multinomial
glbMdlFamilies[["All.X"]] <- c("glmnet")
} else {
glbMdlFamilies[["All.X"]] <- c("glmnet", "glm")
}
glbMdlFamilies[["All.X.Inc"]] <- c("glmnet")
glbMdlFamilies[["RFE.X"]] <- c("glmnet")
# Check if interaction features make RFE better
# glbMdlFamilies[["CSM.X"]] <- setdiff(glbMdlMethods, c("lda", "lda2")) # crashing due to category:.clusterid ??? #c("glmnet", "glm") # non-NULL list is mandatory
# glb_mdl_feats_lst[["CSM.X"]] <- c(NULL
# , <comma-separated-features-vector>
# )
# dAFeats.CSM.X %<d-% c(NULL
# # Interaction feats up to varImp(RFE.X.glmnet) >= 50
# , <comma-separated-features-vector>
# , setdiff(myextract_actual_feats(predictors(glbRFEResults)), c(NULL
# , <comma-separated-features-vector>
# ))
# )
# glb_mdl_feats_lst[["CSM.X"]] <- "%<d-% dAFeats.CSM.X"
glbMdlFamilies[["Final"]] <- c(NULL) # NULL vector acceptable # c("glmnet", "glm")
glbMdlAllowParallel <- list()
#glbMdlAllowParallel[["Final##rcv#glmnet"]] <- FALSE
glbMdlAllowParallel[["All.X##rcv#glm"]] <- FALSE
glbMdlAllowParallel[["All.X#zv.pca#rcv#glmnet"]] <- FALSE
glbMdlAllowParallel[["All.X#ica#rcv#glmnet"]] <- FALSE
glbMdlAllowParallel[["RFE.X#zv.pca#rcv#glmnet"]] <- FALSE
glbMdlAllowParallel[["RFE.X#ica#rcv#glmnet"]] <- FALSE
glbMdlAllowParallel[["RFE.X#nzv.pca.spatialSign#rcv#glmnet"]] <- FALSE
glbMdlAllowParallel[["Final.RFE.X#zv.pca#rcv#glmnet"]] <- FALSE
# Check if tuning parameters make fit better; make it mdlFamily customizable ?
glbMdlTuneParams <- data.frame()
# When glmnet crashes at model$grid with error: ???
glmnetTuneParams <- rbind(data.frame()
,data.frame(parameter = "alpha", vals = "0.100 0.325 0.550 0.775 1.000")
,data.frame(parameter = "lambda", vals = "9.342e-02")
)
# glbMdlTuneParams <- myrbind_df(glbMdlTuneParams,
# cbind(data.frame(mdlId = "<mdlId>"),
# glmnetTuneParams))
#avNNet
# size=[1] 3 5 7 9; decay=[0] 1e-04 0.001 0.01 0.1; bag=[FALSE]; RMSE=1.3300906
#bagEarth
# degree=1 [2] 3; nprune=64 128 256 512 [1024]; RMSE=0.6486663 (up)
# glbMdlTuneParams <- myrbind_df(glbMdlTuneParams, rbind(data.frame()
# ,data.frame(method = "bagEarth", parameter = "nprune", vals = "256")
# ,data.frame(method = "bagEarth", parameter = "degree", vals = "2")
# ))
#earth
# degree=[1]; nprune=2 [9] 17 25 33; RMSE=0.1334478
#gbm
# shrinkage=0.05 [0.10] 0.15 0.20 0.25; n.trees=100 150 200 [250] 300; interaction.depth=[1] 2 3 4 5; n.minobsinnode=[10]; RMSE=0.2008313
# glbMdlTuneParams <- myrbind_df(glbMdlTuneParams, rbind(data.frame()
# ,data.frame(method = "gbm", parameter = "shrinkage", min = 0.05, max = 0.25, by = 0.05)
# ,data.frame(method = "gbm", parameter = "n.trees", min = 100, max = 300, by = 50)
# ,data.frame(method = "gbm", parameter = "interaction.depth", min = 1, max = 5, by = 1)
# ,data.frame(method = "gbm", parameter = "n.minobsinnode", min = 10, max = 10, by = 10)
# #seq(from=0.05, to=0.25, by=0.05)
# ))
#glmnet
# alpha=0.100 [0.325] 0.550 0.775 1.000; lambda=0.0005232693 0.0024288010 0.0112734954 [0.0523269304] 0.2428800957; RMSE=0.6164891
# glbMdlTuneParams <- myrbind_df(glbMdlTuneParams, rbind(data.frame()
# ,data.frame(method = "glmnet", parameter = "alpha", vals = "0.550 0.775 0.8875 0.94375 1.000")
# ,data.frame(method = "glmnet", parameter = "lambda", vals = "9.858855e-05 0.0001971771 0.0009152152 0.0042480525 0.0197177130")
# ))
#nnet
# size=3 5 [7] 9 11; decay=0.0001 0.001 0.01 [0.1] 0.2; RMSE=0.9287422
# glbMdlTuneParams <- myrbind_df(glbMdlTuneParams, rbind(data.frame()
# ,data.frame(method = "nnet", parameter = "size", vals = "3 5 7 9 11")
# ,data.frame(method = "nnet", parameter = "decay", vals = "0.0001 0.0010 0.0100 0.1000 0.2000")
# ))
#rf # Don't bother; results are not deterministic
# mtry=2 35 68 [101] 134; RMSE=0.1339974
# glbMdlTuneParams <- myrbind_df(glbMdlTuneParams, rbind(data.frame()
# ,data.frame(method = "rf", parameter = "mtry", vals = "2 5 9 13 17")
# ))
#rpart
# cp=0.020 [0.025] 0.030 0.035 0.040; RMSE=0.1770237
# glbMdlTuneParams <- myrbind_df(glbMdlTuneParams, rbind(data.frame()
# ,data.frame(method = "rpart", parameter = "cp", vals = "0.004347826 0.008695652 0.017391304 0.021739130 0.034782609")
# ))
#svmLinear
# C=0.01 0.05 [0.10] 0.50 1.00 2.00 3.00 4.00; RMSE=0.1271318; 0.1296718
# glbMdlTuneParams <- myrbind_df(glbMdlTuneParams, rbind(data.frame()
# ,data.frame(method = "svmLinear", parameter = "C", vals = "0.01 0.05 0.1 0.5 1")
# ))
#svmLinear2
# cost=0.0625 0.1250 [0.25] 0.50 1.00; RMSE=0.1276354
# glbMdlTuneParams <- myrbind_df(glbMdlTuneParams, rbind(data.frame()
# ,data.frame(method = "svmLinear2", parameter = "cost", vals = "0.0625 0.125 0.25 0.5 1")
# ))
#svmPoly
# degree=[1] 2 3 4 5; scale=0.01 0.05 [0.1] 0.5 1; C=0.50 1.00 [2.00] 3.00 4.00; RMSE=0.1276130
# glbMdlTuneParams <- myrbind_df(glbMdlTuneParams, rbind(data.frame()
# ,data.frame(method="svmPoly", parameter="degree", min=1, max=5, by=1) #seq(1, 5, 1)
# ,data.frame(method="svmPoly", parameter="scale", vals="0.01, 0.05, 0.1, 0.5, 1")
# ,data.frame(method="svmPoly", parameter="C", vals="0.50, 1.00, 2.00, 3.00, 4.00")
# ))
#svmRadial
# sigma=[0.08674323]; C=0.25 0.50 1.00 [2.00] 4.00; RMSE=0.1614957
#glb2Sav(); all.equal(sav_models_df, glb_models_df)
pkgPreprocMethods <-
# caret version: 6.0.068 # packageVersion("caret")
# operations are applied in this order: zero-variance filter, near-zero variance filter, Box-Cox/Yeo-Johnson/exponential transformation, centering, scaling, range, imputation, PCA, ICA then spatial sign
# *Impute methods needed only if NAs are fed to myfit_mdl
# Also, ordered.factor in caret creates features as Edn.fctr^4 which is treated as an exponent by bagImpute
c(NULL
,"zv", "nzv"
,"BoxCox", "YeoJohnson", "expoTrans"
,"center", "scale", "center.scale", "range"
,"knnImpute", "bagImpute", "medianImpute"
,"zv.pca", "ica", "spatialSign"
,"conditionalX")
glbMdlPreprocMethods <- list( # NULL # : default
"All.X" = list("glmnet" = union(setdiff(pkgPreprocMethods,
c("knnImpute", "bagImpute", "medianImpute")),
# NULL))
c("nzv.spatialSign")))
)
glbMdlPreprocMethods[["RFE.X"]] <- list("glmnet" = union(unlist(glbMdlPreprocMethods[["All.X"]]),
"nzv.pca.spatialSign"))
# Baseline prediction model feature(s)
glb_Baseline_mdl_var <- NULL # or c("<feat>")
glbMdlMetric_terms <- NULL # or matrix(c(
# 0,1,2,3,4,
# 2,0,1,2,3,
# 4,2,0,1,2,
# 6,4,2,0,1,
# 8,6,4,2,0
# ), byrow=TRUE, nrow=5)
glbMdlMetricSummary <- NULL # or "<metric_name>"
glbMdlMetricMaximize <- NULL # or FALSE (TRUE is not the default for both classification & regression)
glbMdlMetricSummaryFn <- NULL # or function(data, lev=NULL, model=NULL) {
# confusion_mtrx <- t(as.matrix(confusionMatrix(data$pred, data$obs)))
# #print(confusion_mtrx)
# #print(confusion_mtrx * glbMdlMetric_terms)
# metric <- sum(confusion_mtrx * glbMdlMetric_terms) / nrow(data)
# names(metric) <- glbMdlMetricSummary
# return(metric)
# }
glbMdlCheckRcv <- FALSE # Turn it on when needed; otherwise takes long time
glb_rcv_n_folds <- 3 # or NULL
glb_rcv_n_repeats <- 3 # or NULL
glb_clf_proba_threshold <- NULL # 0.5
# Model selection criteria
if (glb_is_regression)
glbMdlMetricsEval <- c("min.RMSE.OOB", "max.R.sq.OOB", "max.Adj.R.sq.fit", "min.RMSE.fit")
#glbMdlMetricsEval <- c("min.RMSE.fit", "max.R.sq.fit", "max.Adj.R.sq.fit")
if (glb_is_classification) {
if (glb_is_binomial)
glbMdlMetricsEval <-
c("max.Accuracy.OOB", "max.AUCROCR.OOB", "max.AUCpROC.OOB", "min.aic.fit", "max.Accuracy.fit") else
glbMdlMetricsEval <- c("max.Accuracy.OOB", "max.Kappa.OOB")
}
# select from NULL [no ensemble models], "auto" [all models better than MFO or Baseline], c(mdl_ids in glb_models_lst) [Typically top-rated models in auto]
glb_mdl_ensemble <- NULL
# "%<d-% setdiff(mygetEnsembleAutoMdlIds(), 'CSM.X.rf')"
# c(<comma-separated-mdlIds>
# )
# Only for classifications; for regressions remove "(.*)\\.prob" form the regex
# tmp_fitobs_df <- glbObsFit[, grep(paste0("^", gsub(".", "\\.", mygetPredictIds$value, fixed = TRUE), "CSM\\.X\\.(.*)\\.prob"), names(glbObsFit), value = TRUE)]; cor_mtrx <- cor(tmp_fitobs_df); cor_vctr <- sort(cor_mtrx[row.names(orderBy(~-Overall, varImp(glb_models_lst[["Ensemble.repeatedcv.glmnet"]])$imp))[1], ]); summary(cor_vctr); cor_vctr
#ntv.glm <- glm(reformulate(indepVar, glb_rsp_var), family = "binomial", data = glbObsFit)
#step.glm <- step(ntv.glm)
glbMdlSelId <- NULL #select from c(NULL, "All.X##rcv#glmnet", "RFE.X##rcv#glmnet", <mdlId>)
glbMdlFinId <- NULL #select from c(NULL, glbMdlSelId)
glb_dsp_cols <- c(".pos", glbFeatsId, glbFeatsCategory, glb_rsp_var
# List critical cols excl. above
)
# Output specs
# lclgetfltout_df <- function(obsOutFinDf) {
# require(tidyr)
# obsOutFinDf <- obsOutFinDf %>%
# tidyr::separate("ImageId.x.y", c(".src", ".pos", "x", "y"),
# sep = "#", remove = TRUE, extra = "merge")
# # mnm prefix stands for max_n_mean
# mnmout_df <- obsOutFinDf %>%
# dplyr::group_by(.pos) %>%
# #dplyr::top_n(1, Probability1) %>% # Score = 3.9426
# #dplyr::top_n(2, Probability1) %>% # Score = ???; weighted = 3.94254;
# #dplyr::top_n(3, Probability1) %>% # Score = 3.9418; weighted = 3.94169;
# dplyr::top_n(4, Probability1) %>% # Score = ???; weighted = 3.94149;
# #dplyr::top_n(5, Probability1) %>% # Score = 3.9421; weighted = 3.94178
#
# # dplyr::summarize(xMeanN = mean(as.numeric(x)), yMeanN = mean(as.numeric(y)))
# # dplyr::summarize(xMeanN = weighted.mean(as.numeric(x), Probability1), yMeanN = mean(as.numeric(y)))
# # dplyr::summarize(xMeanN = weighted.mean(as.numeric(x), c(Probability1, 0.2357323, 0.2336925)), yMeanN = mean(as.numeric(y)))
# # dplyr::summarize(xMeanN = weighted.mean(as.numeric(x), c(Probability1)), yMeanN = mean(as.numeric(y)))
# dplyr::summarize(xMeanN = weighted.mean(as.numeric(x), c(Probability1)),
# yMeanN = weighted.mean(as.numeric(y), c(Probability1)))
#
# maxout_df <- obsOutFinDf %>%
# dplyr::group_by(.pos) %>%
# dplyr::summarize(maxProb1 = max(Probability1))
# fltout_df <- merge(maxout_df, obsOutFinDf,
# by.x = c(".pos", "maxProb1"), by.y = c(".pos", "Probability1"),
# all.x = TRUE)
# fmnout_df <- merge(fltout_df, mnmout_df,
# by.x = c(".pos"), by.y = c(".pos"),
# all.x = TRUE)
# return(fmnout_df)
# }
glbObsOut <- list(NULL
# glbFeatsId will be the first output column, by default
,vars = list()
# ,mapFn = function(obsOutFinDf) {
# }
)
#obsOutFinDf <- savobsOutFinDf
# glbObsOut$mapFn <- function(obsOutFinDf) {
# txfout_df <- dplyr::select(obsOutFinDf, -.pos.y) %>%
# dplyr::mutate(
# lunch = levels(glbObsTrn[, "lunch" ])[
# round(mean(as.numeric(glbObsTrn[, "lunch" ])), 0)],
# dinner = levels(glbObsTrn[, "dinner" ])[
# round(mean(as.numeric(glbObsTrn[, "dinner" ])), 0)],
# reserve = levels(glbObsTrn[, "reserve" ])[
# round(mean(as.numeric(glbObsTrn[, "reserve" ])), 0)],
# outdoor = levels(glbObsTrn[, "outdoor" ])[
# round(mean(as.numeric(glbObsTrn[, "outdoor" ])), 0)],
# expensive = levels(glbObsTrn[, "expensive"])[
# round(mean(as.numeric(glbObsTrn[, "expensive"])), 0)],
# liquor = levels(glbObsTrn[, "liquor" ])[
# round(mean(as.numeric(glbObsTrn[, "liquor" ])), 0)],
# table = levels(glbObsTrn[, "table" ])[
# round(mean(as.numeric(glbObsTrn[, "table" ])), 0)],
# classy = levels(glbObsTrn[, "classy" ])[
# round(mean(as.numeric(glbObsTrn[, "classy" ])), 0)],
# kids = levels(glbObsTrn[, "kids" ])[
# round(mean(as.numeric(glbObsTrn[, "kids" ])), 0)]
# )
#
# print("ObsNew output class tables:")
# print(sapply(c("lunch","dinner","reserve","outdoor",
# "expensive","liquor","table",
# "classy","kids"),
# function(feat) table(txfout_df[, feat], useNA = "ifany")))
#
# txfout_df <- txfout_df %>%
# dplyr::mutate(labels = "") %>%
# dplyr::mutate(labels =
# ifelse(lunch != "-1", paste(labels, lunch ), labels)) %>%
# dplyr::mutate(labels =
# ifelse(dinner != "-1", paste(labels, dinner ), labels)) %>%
# dplyr::mutate(labels =
# ifelse(reserve != "-1", paste(labels, reserve ), labels)) %>%
# dplyr::mutate(labels =
# ifelse(outdoor != "-1", paste(labels, outdoor ), labels)) %>%
# dplyr::mutate(labels =
# ifelse(expensive != "-1", paste(labels, expensive), labels)) %>%
# dplyr::mutate(labels =
# ifelse(liquor != "-1", paste(labels, liquor ), labels)) %>%
# dplyr::mutate(labels =
# ifelse(table != "-1", paste(labels, table ), labels)) %>%
# dplyr::mutate(labels =
# ifelse(classy != "-1", paste(labels, classy ), labels)) %>%
# dplyr::mutate(labels =
# ifelse(kids != "-1", paste(labels, kids ), labels)) %>%
# dplyr::select(business_id, labels)
# return(txfout_df)
# }
#if (!is.null(glbObsOut$mapFn)) obsOutFinDf <- glbObsOut$mapFn(obsOutFinDf); print(head(obsOutFinDf))
glb_out_obs <- NULL # select from c(NULL : default to "new", "all", "new", "trn")
if (glb_is_classification && glb_is_binomial) {
# glbObsOut$vars[["Probability1"]] <-
# "%<d-% glbObsNew[, mygetPredictIds(glb_rsp_var, glbMdlFinId)$prob]"
# glbObsOut$vars[[glb_rsp_var_raw]] <-
# "%<d-% glb_map_rsp_var_to_raw(glbObsNew[,
# mygetPredictIds(glb_rsp_var, glbMdlFinId)$value])"
glbObsOut$vars[["Predictions"]] <-
"%<d-% glb_map_rsp_var_to_raw(glbObsNew[,
mygetPredictIds(glb_rsp_var, glbMdlFinId)$value])"
} else {
# glbObsOut$vars[[glbFeatsId]] <-
# "%<d-% as.integer(gsub('Test#', '', glbObsNew[, glbFeatsId]))"
glbObsOut$vars[[glb_rsp_var]] <-
"%<d-% glbObsNew[, mygetPredictIds(glb_rsp_var, glbMdlFinId)$value]"
# for (outVar in setdiff(glbFeatsExcludeLcl, glb_rsp_var_raw))
# glbObsOut$vars[[outVar]] <-
# paste0("%<d-% mean(glbObsAll[, \"", outVar, "\"], na.rm = TRUE)")
}
# glbObsOut$vars[[glb_rsp_var_raw]] <- glb_rsp_var_raw
# glbObsOut$vars[[paste0(head(unlist(strsplit(mygetPredictIds$value, "")), -1), collapse = "")]] <-
glbOutStackFnames <- NULL #: default
# c("ebayipads_txt_assoc1_out_bid1_stack.csv") # manual stack
# c("ebayipads_finmdl_bid1_out_nnet_1.csv") # universal stack
glbOut <- list(pfx = "Votes_RFEX_cnk04_rest_")
# lclImageSampleSeed <- 129
glbOutDataVizFname <- NULL # choose from c(NULL, "<projectId>_obsall.csv")
glbChunks <- list(labels = c("set_global_options_wd","set_global_options"
,"import.data","inspect.data","scrub.data","transform.data"
,"extract.features"
,"extract.features.datetime","extract.features.image","extract.features.price"
,"extract.features.text","extract.features.string"
,"extract.features.end"
,"manage.missing.data","cluster.data","partition.data.training","select.features"
,"fit.models_0","fit.models_1","fit.models_2","fit.models_3"
,"fit.data.training_0","fit.data.training_1"
,"predict.data.new"
,"display.session.info"))
# To ensure that all chunks in this script are in glbChunks
if (!is.null(chkChunksLabels <- knitr::all_labels()) && # knitr::all_labels() doesn't work in console runs
!identical(chkChunksLabels, glbChunks$labels)) {
print(sprintf("setdiff(chkChunksLabels, glbChunks$labels): %s",
setdiff(chkChunksLabels, glbChunks$labels)))
print(sprintf("setdiff(glbChunks$labels, chkChunksLabels): %s",
setdiff(glbChunks$labels, chkChunksLabels)))
}
glbChunks[["first"]] <- "fit.models_1" #default: script will load envir from previous chunk
glbChunks[["last" ]] <- NULL #default: script will save envir at end of this chunk
glbChunks[["inpFilePathName"]] <- "data/Votes_RFEX_cnk04_fit_models_1_fit.models_1.RData" # "data/<prvScriptName>_<lstChunkLbl>.RData"
#mysavChunk(glbOut$pfx, glbChunks[["last"]]) # called from myevlChunk
# Temporary: Delete this function (if any) from here after appropriate .RData file is saved
# glbgetModelSelectFormula <- function() {
# model_evl_terms <- c(NULL)
# # min.aic.fit might not be avl
# lclMdlEvlCriteria <-
# glbMdlMetricsEval[glbMdlMetricsEval %in% names(glb_models_df)]
# for (metric in lclMdlEvlCriteria)
# model_evl_terms <- c(model_evl_terms,
# ifelse(length(grep("max", metric)) > 0, "-", "+"), metric)
# if (glb_is_classification && glb_is_binomial)
# model_evl_terms <- c(model_evl_terms, "-", "opt.prob.threshold.OOB")
# model_sel_frmla <- as.formula(paste(c("~ ", model_evl_terms), collapse = " "))
# return(model_sel_frmla)
# }
# glbgetDisplayModelsDf <- function() {
# dsp_models_cols <- c("id",
# glbMdlMetricsEval[glbMdlMetricsEval %in% names(glb_models_df)],
# grep("opt.", names(glb_models_df), fixed = TRUE, value = TRUE))
# dsp_models_df <-
# #orderBy(glbgetModelSelectFormula(), glb_models_df)[, c("id", glbMdlMetricsEval)]
# orderBy(glbgetModelSelectFormula(), glb_models_df)[, dsp_models_cols]
# nCvMdl <- sapply(glb_models_lst, function(mdl) nrow(mdl$results))
# nParams <- sapply(glb_models_lst, function(mdl) ifelse(mdl$method == "custom", 0,
# nrow(subset(modelLookup(mdl$method), parameter != "parameter"))))
#
# # nCvMdl <- nCvMdl[names(nCvMdl) != "avNNet"]
# # nParams <- nParams[names(nParams) != "avNNet"]
#
# if (length(cvMdlProblems <- nCvMdl[nCvMdl <= nParams]) > 0) {
# print("Cross Validation issues:")
# warning("Cross Validation issues:")
# print(cvMdlProblems)
# }
#
# pltMdls <- setdiff(names(nCvMdl), names(cvMdlProblems))
# pltMdls <- setdiff(pltMdls, names(nParams[nParams == 0]))
#
# # length(pltMdls) == 21
# png(paste0(glbOut$pfx, "bestTune.png"), width = 480 * 2, height = 480 * 4)
# grid.newpage()
# pushViewport(viewport(layout = grid.layout(ceiling(length(pltMdls) / 2.0), 2)))
# pltIx <- 1
# for (mdlId in pltMdls) {
# print(ggplot(glb_models_lst[[mdlId]], highBestTune = TRUE) + labs(title = mdlId),
# vp = viewport(layout.pos.row = ceiling(pltIx / 2.0),
# layout.pos.col = ((pltIx - 1) %% 2) + 1))
# pltIx <- pltIx + 1
# }
# dev.off()
#
# if (all(row.names(dsp_models_df) != dsp_models_df$id))
# row.names(dsp_models_df) <- dsp_models_df$id
# return(dsp_models_df)
# }
# Inspect max OOB FP
#chkObsOOB <- subset(glbObsOOB, !label.fctr.All.X..rcv.glmnet.is.acc)
#chkObsOOBFP <- subset(chkObsOOB, label.fctr.All.X..rcv.glmnet == "left_eye_center") %>% dplyr::mutate(Probability1 = label.fctr.All.X..rcv.glmnet.prob) %>% select(-.src, -.pos, -x, -y) %>% lclgetfltout_df() %>% mutate(obj.distance = (((as.numeric(x) - left_eye_center_x.int) ^ 2) + ((as.numeric(y) - left_eye_center_y.int) ^ 2)) ^ 0.5) %>% dplyr::top_n(5, obj.distance) %>% dplyr::top_n(5, -patch.cor)
#
#newImgObs <- glbObsNew[(glbObsNew$ImageId == "Test#0001"), ]; print(newImgObs[which.max(newImgObs$label.fctr.Final..rcv.glmnet.prob), ])
#OOBImgObs <- glbObsOOB[(glbObsOOB$ImageId == "Train#0003"), ]; print(OOBImgObs[which.max(OOBImgObs$label.fctr.All.X..rcv.glmnet.prob), ])
#mygetImage(which(glbObsAll[, glbFeatsId] == "Train#0003"), names(glbFeatsImage)[1], plot = TRUE, featHighlight = c("left_eye_center_x", "left_eye_center_y"), ovrlHighlight = c(66, 35))
# Depict process
glb_analytics_pn <- petrinet(name = "glb_analytics_pn",
trans_df = data.frame(id = 1:6,
name = c("data.training.all","data.new",
"model.selected","model.final",
"data.training.all.prediction","data.new.prediction"),
x=c( -5,-5,-15,-25,-25,-35),
y=c( -5, 5, 0, 0, -5, 5)
),
places_df=data.frame(id=1:4,
name=c("bgn","fit.data.training.all","predict.data.new","end"),
x=c( -0, -20, -30, -40),
y=c( 0, 0, 0, 0),
M0=c( 3, 0, 0, 0)
),
arcs_df = data.frame(
begin = c("bgn","bgn","bgn",
"data.training.all","model.selected","fit.data.training.all",
"fit.data.training.all","model.final",
"data.new","predict.data.new",
"data.training.all.prediction","data.new.prediction"),
end = c("data.training.all","data.new","model.selected",
"fit.data.training.all","fit.data.training.all","model.final",
"data.training.all.prediction","predict.data.new",
"predict.data.new","data.new.prediction",
"end","end")
))
#print(ggplot.petrinet(glb_analytics_pn))
print(ggplot.petrinet(glb_analytics_pn) + coord_flip())## Loading required package: grid
glb_analytics_avl_objs <- NULL
glb_chunks_df <- myadd_chunk(NULL,
ifelse(is.null(glbChunks$first), "import.data", glbChunks$first))## label step_major step_minor label_minor bgn end elapsed
## 1 fit.models_1 1 0 0 7.72 NA NAStep 1.0: fit models_1
chunk option: eval=
Step 1.0: fit models_1
Step 1.0: fit models_1
Step 1.0: fit models_1
Step 1.0: fit models_1
Step 1.0: fit models_1
Step 1.0: fit models_1
```{r extract.features.image, cache=FALSE, echo=FALSE, fig.height=5, fig.width=5, eval=myevlChunk(glbChunks, glbOut$pfx)}
Step 1.0: fit models_1
Step 1.0: fit models_1
Step 1.0: fit models_1
Step 1.0: fit models_1
Step 1.0: fit models_1
Step 1.0: fit models_1
```{r cluster.data, cache=FALSE, echo=FALSE, eval=myevlChunk(glbChunks, glbOut$pfx)}
Step 1.0: fit models_1
Step 1.0: fit models_1
```{r select.features, cache=FALSE, echo=FALSE, eval=myevlChunk(glbChunks, glbOut$pfx)}
Step 1.0: fit models_1
fit.models_0_chunk_df <- myadd_chunk(NULL, "fit.models_0_bgn", label.minor = "setup")
# load(paste0(glbOut$pfx, "dsk.RData"))
glbgetModelSelectFormula <- function() {
model_evl_terms <- c(NULL)
# min.aic.fit might not be avl
lclMdlEvlCriteria <-
glbMdlMetricsEval[glbMdlMetricsEval %in% names(glb_models_df)]
for (metric in lclMdlEvlCriteria)
model_evl_terms <- c(model_evl_terms,
ifelse(length(grep("max", metric)) > 0, "-", "+"), metric)
if (glb_is_classification && glb_is_binomial)
model_evl_terms <- c(model_evl_terms, "-", "opt.prob.threshold.OOB")
model_sel_frmla <- as.formula(paste(c("~ ", model_evl_terms), collapse = " "))
return(model_sel_frmla)
}
glbgetDisplayModelsDf <- function() {
dsp_models_cols <- c("id",
glbMdlMetricsEval[glbMdlMetricsEval %in% names(glb_models_df)],
grep("opt.", names(glb_models_df), fixed = TRUE, value = TRUE))
dsp_models_df <-
#orderBy(glbgetModelSelectFormula(), glb_models_df)[, c("id", glbMdlMetricsEval)]
orderBy(glbgetModelSelectFormula(), glb_models_df)[, dsp_models_cols]
nCvMdl <- sapply(glb_models_lst, function(mdl) nrow(mdl$results))
nParams <- sapply(glb_models_lst, function(mdl) ifelse(mdl$method == "custom", 0,
nrow(subset(modelLookup(mdl$method), parameter != "parameter"))))
# nCvMdl <- nCvMdl[names(nCvMdl) != "avNNet"]
# nParams <- nParams[names(nParams) != "avNNet"]
if (length(cvMdlProblems <- nCvMdl[nCvMdl <= nParams]) > 0) {
print("Cross Validation issues:")
warning("Cross Validation issues:")
print(cvMdlProblems)
}
pltMdls <- setdiff(names(nCvMdl), names(cvMdlProblems))
pltMdls <- setdiff(pltMdls, names(nParams[nParams == 0]))
# length(pltMdls) == 21
png(paste0(glbOut$pfx, "bestTune.png"), width = 480 * 2, height = 480 * 4)
grid.newpage()
pushViewport(viewport(layout = grid.layout(ceiling(length(pltMdls) / 2.0), 2)))
pltIx <- 1
for (mdlId in pltMdls) {
print(ggplot(glb_models_lst[[mdlId]], highBestTune = TRUE) + labs(title = mdlId),
vp = viewport(layout.pos.row = ceiling(pltIx / 2.0),
layout.pos.col = ((pltIx - 1) %% 2) + 1))
pltIx <- pltIx + 1
}
dev.off()
if (all(row.names(dsp_models_df) != dsp_models_df$id))
row.names(dsp_models_df) <- dsp_models_df$id
return(dsp_models_df)
}
#glbgetDisplayModelsDf()
if (glb_is_classification && glb_is_binomial &&
(length(unique(glbObsFit[, glb_rsp_var])) < 2))
stop("glbObsFit$", glb_rsp_var, ": contains less than 2 unique values: ",
paste0(unique(glbObsFit[, glb_rsp_var]), collapse=", "))
max_cor_y_x_vars <- orderBy(~ -cor.y.abs,
subset(glb_feats_df, (exclude.as.feat == 0) & !nzv & !is.cor.y.abs.low &
is.na(cor.high.X)))[1:2, "id"]
max_cor_y_x_vars <- max_cor_y_x_vars[!is.na(max_cor_y_x_vars)]
if (length(max_cor_y_x_vars) < 2)
max_cor_y_x_vars <- union(max_cor_y_x_vars, ".pos")
if (!is.null(glb_Baseline_mdl_var)) {
if ((max_cor_y_x_vars[1] != glb_Baseline_mdl_var) &
(glb_feats_df[glb_feats_df$id == max_cor_y_x_vars[1], "cor.y.abs"] >
glb_feats_df[glb_feats_df$id == glb_Baseline_mdl_var, "cor.y.abs"]))
stop(max_cor_y_x_vars[1], " has a higher correlation with ", glb_rsp_var,
" than the Baseline var: ", glb_Baseline_mdl_var)
}
glb_model_type <- ifelse(glb_is_regression, "regression", "classification")
# Model specs
# c("id.prefix", "method", "type",
# # trainControl params
# "preProc.method", "cv.n.folds", "cv.n.repeats", "summary.fn",
# # train params
# "metric", "metric.maximize", "tune.df")
# Baseline
if (!is.null(glb_Baseline_mdl_var)) {
fit.models_0_chunk_df <- myadd_chunk(fit.models_0_chunk_df,
paste0("fit.models_0_", "Baseline"), major.inc = FALSE,
label.minor = "mybaseln_classfr")
ret_lst <- myfit_mdl(mdl_id="Baseline",
model_method="mybaseln_classfr",
indepVar=glb_Baseline_mdl_var,
rsp_var=glb_rsp_var,
fit_df=glbObsFit, OOB_df=glbObsOOB)
}
# Most Frequent Outcome "MFO" model: mean(y) for regression
# Not using caret's nullModel since model stats not avl
# Cannot use rpart for multinomial classification since it predicts non-MFO
if (glb_is_classification) {
fit.models_0_chunk_df <- myadd_chunk(fit.models_0_chunk_df,
paste0("fit.models_0_", "MFO"), major.inc = FALSE,
label.minor = "myMFO_classfr")
ret_lst <- myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = "MFO", type = glb_model_type, trainControl.method = "none",
train.method = ifelse(glb_is_regression, "lm", "myMFO_classfr"))),
indepVar = ".rnorm", rsp_var = glb_rsp_var,
fit_df = glbObsFit, OOB_df = glbObsOOB)
# "random" model - only for classification;
# none needed for regression since it is same as MFO
fit.models_0_chunk_df <- myadd_chunk(fit.models_0_chunk_df,
paste0("fit.models_0_", "Random"), major.inc = FALSE,
label.minor = "myrandom_classfr")
#stop(here"); glb2Sav(); all.equal(glb_models_df, sav_models_df)
ret_lst <- myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = "Random", type = glb_model_type, trainControl.method = "none",
train.method = "myrandom_classfr")),
indepVar = ".rnorm", rsp_var = glb_rsp_var,
fit_df = glbObsFit, OOB_df = glbObsOOB)
}
# Max.cor.Y
# Check impact of cv
# rpart is not a good candidate since caret does not optimize cp (only tuning parameter of rpart) well
fit.models_0_chunk_df <- myadd_chunk(fit.models_0_chunk_df,
paste0("fit.models_0_", "Max.cor.Y.rcv.*X*"), major.inc = FALSE,
label.minor = "glmnet")
ret_lst <- myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = "Max.cor.Y.rcv.1X1", type = glb_model_type, trainControl.method = "none",
train.method = "glmnet")),
indepVar = max_cor_y_x_vars, rsp_var = glb_rsp_var,
fit_df = glbObsFit, OOB_df = glbObsOOB)
if (glbMdlCheckRcv) {
# rcv_n_folds == 1 & rcv_n_repeats > 1 crashes
for (rcv_n_folds in seq(3, glb_rcv_n_folds + 2, 2))
for (rcv_n_repeats in seq(1, glb_rcv_n_repeats + 2, 2)) {
# Experiment specific code to avoid caret crash
# lcl_tune_models_df <- rbind(data.frame()
# ,data.frame(method = "glmnet", parameter = "alpha",
# vals = "0.100 0.325 0.550 0.775 1.000")
# ,data.frame(method = "glmnet", parameter = "lambda",
# vals = "9.342e-02")
# )
ret_lst <- myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst =
list(
id.prefix = paste0("Max.cor.Y.rcv.", rcv_n_folds, "X", rcv_n_repeats),
type = glb_model_type,
# tune.df = lcl_tune_models_df,
trainControl.method = "repeatedcv",
trainControl.number = rcv_n_folds,
trainControl.repeats = rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
train.method = "glmnet", train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize)),
indepVar = max_cor_y_x_vars, rsp_var = glb_rsp_var,
fit_df = glbObsFit, OOB_df = glbObsOOB)
}
# Add parallel coordinates graph of glb_models_df[, glbMdlMetricsEval] to evaluate cv parameters
tmp_models_cols <- c("id", "max.nTuningRuns",
glbMdlMetricsEval[glbMdlMetricsEval %in% names(glb_models_df)],
grep("opt.", names(glb_models_df), fixed = TRUE, value = TRUE))
print(myplot_parcoord(obs_df = subset(glb_models_df,
grepl("Max.cor.Y.rcv.", id, fixed = TRUE),
select = -feats)[, tmp_models_cols],
id_var = "id"))
}
# Useful for stacking decisions
# fit.models_0_chunk_df <- myadd_chunk(fit.models_0_chunk_df,
# paste0("fit.models_0_", "Max.cor.Y[rcv.1X1.cp.0|]"), major.inc = FALSE,
# label.minor = "rpart")
#
# ret_lst <- myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
# id.prefix = "Max.cor.Y.rcv.1X1.cp.0", type = glb_model_type, trainControl.method = "none",
# train.method = "rpart",
# tune.df=data.frame(method="rpart", parameter="cp", min=0.0, max=0.0, by=0.1))),
# indepVar=max_cor_y_x_vars, rsp_var=glb_rsp_var,
# fit_df=glbObsFit, OOB_df=glbObsOOB)
#stop(here"); glb2Sav(); all.equal(glb_models_df, sav_models_df)
# if (glb_is_regression || glb_is_binomial) # For multinomials this model will be run next by default
ret_lst <- myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = "Max.cor.Y",
type = glb_model_type, trainControl.method = "repeatedcv",
trainControl.number = glb_rcv_n_folds,
trainControl.repeats = glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
trainControl.allowParallel = glbMdlAllowParallel,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method = "rpart")),
indepVar = max_cor_y_x_vars, rsp_var = glb_rsp_var,
fit_df = glbObsFit, OOB_df = glbObsOOB)
if ((length(glbFeatsDateTime) > 0) &&
(sum(grepl(paste(names(glbFeatsDateTime), "\\.day\\.minutes\\.poly\\.", sep = ""),
names(glbObsAll))) > 0)) {
fit.models_0_chunk_df <- myadd_chunk(fit.models_0_chunk_df,
paste0("fit.models_0_", "Max.cor.Y.Time.Poly"), major.inc = FALSE,
label.minor = "glmnet")
indepVars <- c(max_cor_y_x_vars,
grep(paste(names(glbFeatsDateTime), "\\.day\\.minutes\\.poly\\.", sep = ""),
names(glbObsAll), value = TRUE))
indepVars <- myadjustInteractionFeats(glb_feats_df, indepVars)
ret_lst <- myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = "Max.cor.Y.Time.Poly",
type = glb_model_type, trainControl.method = "repeatedcv",
trainControl.number = glb_rcv_n_folds, trainControl.repeats = glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
trainControl.allowParallel = glbMdlAllowParallel,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method = "glmnet")),
indepVar = indepVars,
rsp_var = glb_rsp_var,
fit_df = glbObsFit, OOB_df = glbObsOOB)
}
if ((length(glbFeatsDateTime) > 0) &&
(sum(grepl(paste(names(glbFeatsDateTime), "\\.last[[:digit:]]", sep = ""),
names(glbObsAll))) > 0)) {
fit.models_0_chunk_df <- myadd_chunk(fit.models_0_chunk_df,
paste0("fit.models_0_", "Max.cor.Y.Time.Lag"), major.inc = FALSE,
label.minor = "glmnet")
indepVars <- c(max_cor_y_x_vars,
grep(paste(names(glbFeatsDateTime), "\\.last[[:digit:]]", sep = ""),
names(glbObsAll), value = TRUE))
indepVars <- myadjustInteractionFeats(glb_feats_df, indepVars)
ret_lst <- myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = "Max.cor.Y.Time.Lag",
type = glb_model_type,
tune.df = glbMdlTuneParams,
trainControl.method = "repeatedcv",
trainControl.number = glb_rcv_n_folds, trainControl.repeats = glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
trainControl.allowParallel = glbMdlAllowParallel,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method = "glmnet")),
indepVar = indepVars,
rsp_var = glb_rsp_var,
fit_df = glbObsFit, OOB_df = glbObsOOB)
}
if (length(glbFeatsText) > 0) {
fit.models_0_chunk_df <- myadd_chunk(fit.models_0_chunk_df,
paste0("fit.models_0_", "Txt.*"), major.inc = FALSE,
label.minor = "glmnet")
indepVars <- c(max_cor_y_x_vars)
for (txtFeat in names(glbFeatsText))
indepVars <- union(indepVars,
grep(paste(str_to_upper(substr(txtFeat, 1, 1)), "\\.(?!([T|P]\\.))", sep = ""),
names(glbObsAll), perl = TRUE, value = TRUE))
indepVars <- myadjustInteractionFeats(glb_feats_df, indepVars)
ret_lst <- myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = "Max.cor.Y.Text.nonTP",
type = glb_model_type,
tune.df = glbMdlTuneParams,
trainControl.method = "repeatedcv",
trainControl.number = glb_rcv_n_folds, trainControl.repeats = glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
trainControl.allowParallel = glbMdlAllowParallel,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method = "glmnet")),
indepVar = indepVars,
rsp_var = glb_rsp_var,
fit_df = glbObsFit, OOB_df = glbObsOOB)
indepVars <- c(max_cor_y_x_vars)
for (txtFeat in names(glbFeatsText))
indepVars <- union(indepVars,
grep(paste(str_to_upper(substr(txtFeat, 1, 1)), "\\.T\\.", sep = ""),
names(glbObsAll), perl = TRUE, value = TRUE))
indepVars <- myadjustInteractionFeats(glb_feats_df, indepVars)
ret_lst <- myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = "Max.cor.Y.Text.onlyT",
type = glb_model_type,
tune.df = glbMdlTuneParams,
trainControl.method = "repeatedcv",
trainControl.number = glb_rcv_n_folds, trainControl.repeats = glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method = "glmnet")),
indepVar = indepVars,
rsp_var = glb_rsp_var,
fit_df = glbObsFit, OOB_df = glbObsOOB)
indepVars <- c(max_cor_y_x_vars)
for (txtFeat in names(glbFeatsText))
indepVars <- union(indepVars,
grep(paste(str_to_upper(substr(txtFeat, 1, 1)), "\\.P\\.", sep = ""),
names(glbObsAll), perl = TRUE, value = TRUE))
indepVars <- myadjustInteractionFeats(glb_feats_df, indepVars)
ret_lst <- myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = "Max.cor.Y.Text.onlyP",
type = glb_model_type,
tune.df = glbMdlTuneParams,
trainControl.method = "repeatedcv",
trainControl.number = glb_rcv_n_folds, trainControl.repeats = glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
trainControl.allowParallel = glbMdlAllowParallel,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method = "glmnet")),
indepVar = indepVars,
rsp_var = glb_rsp_var,
fit_df = glbObsFit, OOB_df = glbObsOOB)
}
# Interactions.High.cor.Y
if (length(int_feats <- setdiff(setdiff(unique(glb_feats_df$cor.high.X), NA),
subset(glb_feats_df, nzv)$id)) > 0) {
fit.models_0_chunk_df <- myadd_chunk(fit.models_0_chunk_df,
paste0("fit.models_0_", "Interact.High.cor.Y"), major.inc = FALSE,
label.minor = "glmnet")
ret_lst <- myfit_mdl(mdl_specs_lst=myinit_mdl_specs_lst(mdl_specs_lst=list(
id.prefix="Interact.High.cor.Y",
type=glb_model_type, trainControl.method="repeatedcv",
trainControl.number=glb_rcv_n_folds, trainControl.repeats=glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
trainControl.allowParallel = glbMdlAllowParallel,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method="glmnet")),
indepVar=c(max_cor_y_x_vars, paste(max_cor_y_x_vars[1], int_feats, sep=":")),
rsp_var=glb_rsp_var,
fit_df=glbObsFit, OOB_df=glbObsOOB)
}
# Low.cor.X
fit.models_0_chunk_df <- myadd_chunk(fit.models_0_chunk_df,
paste0("fit.models_0_", "Low.cor.X"), major.inc = FALSE,
label.minor = "glmnet")
indepVar <- mygetIndepVar(glb_feats_df)
ret_lst <- myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = "Low.cor.X",
type = glb_model_type,
tune.df = glbMdlTuneParams,
trainControl.method = "repeatedcv",
trainControl.number = glb_rcv_n_folds, trainControl.repeats = glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
trainControl.allowParallel = glbMdlAllowParallel,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method = "glmnet")),
indepVar = indepVar, rsp_var = glb_rsp_var,
fit_df = glbObsFit, OOB_df = glbObsOOB)
fit.models_0_chunk_df <-
myadd_chunk(fit.models_0_chunk_df, "fit.models_0_end", major.inc = FALSE,
label.minor = "teardown")
rm(ret_lst)
glb_chunks_df <- myadd_chunk(glb_chunks_df, "fit.models", major.inc = FALSE)```{r fit.models_1, cache=FALSE, fig.height=10, fig.width=15, eval=myevlChunk(glbChunks, glbOut$pfx)}
## label step_major step_minor label_minor bgn end elapsed
## 1 fit.models_1_bgn 1 0 setup 317.965 NA NA## label step_major step_minor label_minor bgn end
## 1 fit.models_1_bgn 1 0 setup 317.965 317.978
## 2 fit.models_1_All.X 1 1 setup 317.978 NA
## elapsed
## 1 0.013
## 2 NA
## label step_major step_minor label_minor bgn end
## 2 fit.models_1_All.X 1 1 setup 317.978 317.988
## 3 fit.models_1_All.X 1 2 glmnet 317.988 NA
## elapsed
## 2 0.01
## 3 NA
## [1] "skipping fitting model: All.X##rcv#glmnet"
## label step_major step_minor label_minor bgn end
## 3 fit.models_1_All.X 1 2 glmnet 317.988 317.993
## 4 fit.models_1_All.X 1 3 glm 317.994 NA
## elapsed
## 3 0.005
## 4 NA
## [1] "skipping fitting model: All.X##rcv#glm"
## label step_major step_minor label_minor bgn end
## 4 fit.models_1_All.X 1 3 glm 317.994 317.998
## 5 fit.models_1_All.X.Inc 1 4 setup 317.999 NA
## elapsed
## 4 0.004
## 5 NA## Loading required package: glmnet## Loading required package: Matrix## Loaded glmnet 2.0-5## label step_major step_minor label_minor bgn end
## 5 fit.models_1_All.X.Inc 1 4 setup 317.999 318.291
## 6 fit.models_1_All.X.Inc 1 5 glmnet 318.292 NA
## elapsed
## 5 0.292
## 6 NA
## [1] "skipping fitting model: All.X.Inc#nzv.spatialSign#rcv#glmnet"
## label step_major step_minor label_minor bgn end
## 6 fit.models_1_All.X.Inc 1 5 glmnet 318.292 318.301
## 7 fit.models_1_RFE.X 1 6 setup 318.302 NA
## elapsed
## 6 0.009
## 7 NA
## label step_major step_minor label_minor bgn end
## 7 fit.models_1_RFE.X 1 6 setup 318.302 318.308
## 8 fit.models_1_RFE.X 1 7 glmnet 318.308 NA
## elapsed
## 7 0.006
## 8 NA
## [1] "skipping fitting model: RFE.X##rcv#glmnet"## label step_major step_minor label_minor bgn end
## 8 fit.models_1_RFE.X 1 7 glmnet 318.308 318.316
## 9 fit.models_1_preProc 1 8 preProc 318.316 NA
## elapsed
## 8 0.008
## 9 NA## Loading required package: gdata## gdata: read.xls support for 'XLS' (Excel 97-2004) files ENABLED.## ## gdata: read.xls support for 'XLSX' (Excel 2007+) files ENABLED.##
## Attaching package: 'gdata'## The following objects are masked from 'package:dplyr':
##
## combine, first, last## The following object is masked from 'package:stats':
##
## nobs## The following object is masked from 'package:utils':
##
## object.size## [1] "skipping fitting model: All.X#zv#rcv#glmnet"
## [1] "skipping fitting model: All.X#nzv#rcv#glmnet"
## [1] "skipping fitting model: All.X#BoxCox#rcv#glmnet"
## [1] "skipping fitting model: All.X#YeoJohnson#rcv#glmnet"
## [1] "skipping fitting model: All.X#expoTrans#rcv#glmnet"
## [1] "skipping fitting model: All.X#center#rcv#glmnet"
## [1] "skipping fitting model: All.X#scale#rcv#glmnet"
## [1] "skipping fitting model: All.X#center.scale#rcv#glmnet"
## [1] "skipping fitting model: All.X#range#rcv#glmnet"
## [1] "skipping fitting model: All.X#zv.pca#rcv#glmnet"
## [1] "skipping fitting model: All.X#ica#rcv#glmnet"
## [1] "skipping fitting model: All.X#spatialSign#rcv#glmnet"
## [1] "skipping fitting model: All.X#conditionalX#rcv#glmnet"
## [1] "skipping fitting model: All.X#nzv.spatialSign#rcv#glmnet"
## [1] "skipping fitting model: RFE.X#zv#rcv#glmnet"
## [1] "skipping fitting model: RFE.X#nzv#rcv#glmnet"
## [1] "skipping fitting model: RFE.X#BoxCox#rcv#glmnet"
## [1] "skipping fitting model: RFE.X#YeoJohnson#rcv#glmnet"
## [1] "skipping fitting model: RFE.X#expoTrans#rcv#glmnet"
## [1] "skipping fitting model: RFE.X#center#rcv#glmnet"
## [1] "skipping fitting model: RFE.X#scale#rcv#glmnet"
## [1] "skipping fitting model: RFE.X#center.scale#rcv#glmnet"
## [1] "skipping fitting model: RFE.X#range#rcv#glmnet"
## [1] "skipping fitting model: RFE.X#zv.pca#rcv#glmnet"
## [1] "skipping fitting model: RFE.X#ica#rcv#glmnet"
## [1] "skipping fitting model: RFE.X#spatialSign#rcv#glmnet"
## [1] "skipping fitting model: RFE.X#conditionalX#rcv#glmnet"
## [1] "skipping fitting model: RFE.X#nzv.spatialSign#rcv#glmnet"
## [1] "skipping fitting model: RFE.X#nzv.pca.spatialSign#rcv#glmnet"## min.elapsedtime.everything
## Random###myrandom_classfr 0.265
## MFO###myMFO_classfr 0.399
## Max.cor.Y.rcv.1X1###glmnet 0.753
## Max.cor.Y##rcv#rpart 1.624
## Interact.High.cor.Y##rcv#glmnet 5.284
## All.X##rcv#glm 14.312
## RFE.X##rcv#glmnet 21.949
## All.X#conditionalX#rcv#glmnet 22.580
## All.X##rcv#glmnet 22.609
## Low.cor.X##rcv#glmnet 22.775
## RFE.X#conditionalX#rcv#glmnet 23.022
## All.X#zv#rcv#glmnet 23.568
## RFE.X#zv#rcv#glmnet 24.269
## RFE.X#scale#rcv#glmnet 25.649
## RFE.X#center#rcv#glmnet 25.755
## All.X#scale#rcv#glmnet 26.279
## All.X#center#rcv#glmnet 26.891
## RFE.X#range#rcv#glmnet 27.116
## RFE.X#BoxCox#rcv#glmnet 28.258
## All.X#range#rcv#glmnet 28.776
## All.X#BoxCox#rcv#glmnet 29.405
## RFE.X#center.scale#rcv#glmnet 29.657
## All.X#center.scale#rcv#glmnet 30.287
## RFE.X#nzv#rcv#glmnet 34.109
## All.X#nzv#rcv#glmnet 36.115
## All.X#spatialSign#rcv#glmnet 36.885
## RFE.X#ica#rcv#glmnet 39.908
## All.X#ica#rcv#glmnet 41.240
## RFE.X#spatialSign#rcv#glmnet 46.699
## RFE.X#nzv.spatialSign#rcv#glmnet 51.794
## All.X#nzv.spatialSign#rcv#glmnet 52.416
## All.X#YeoJohnson#rcv#glmnet 58.542
## All.X.Inc#nzv.spatialSign#rcv#glmnet 67.172
## All.X#expoTrans#rcv#glmnet 70.569
## RFE.X#expoTrans#rcv#glmnet 71.462
## RFE.X#zv.pca#rcv#glmnet 75.364
## All.X#zv.pca#rcv#glmnet 77.239
## RFE.X#YeoJohnson#rcv#glmnet 84.187
## RFE.X#nzv.pca.spatialSign#rcv#glmnet 533.967## label step_major step_minor label_minor bgn end
## 9 fit.models_1_preProc 1 8 preProc 318.316 319.658
## 10 fit.models_1_end 1 9 teardown 319.658 NA
## elapsed
## 9 1.342
## 10 NA## label step_major step_minor label_minor bgn end elapsed
## 1 fit.models_1 1 0 0 7.720 319.664 311.944
## 2 fit.models 1 1 1 319.665 NA NAfit.models_2_chunk_df <-
myadd_chunk(NULL, "fit.models_2_bgn", label.minor = "setup")## label step_major step_minor label_minor bgn end elapsed
## 1 fit.models_2_bgn 1 0 setup 319.77 NA NArequire(reshape2)## Loading required package: reshape2plt_models_df <- glb_models_df[, -grep("SD|Upper|Lower", names(glb_models_df))]
for (var in grep("^min.", names(plt_models_df), value=TRUE)) {
plt_models_df[, sub("min.", "inv.", var)] <-
#ifelse(all(is.na(tmp <- plt_models_df[, var])), NA, 1.0 / tmp)
1.0 / plt_models_df[, var]
plt_models_df <- plt_models_df[ , -grep(var, names(plt_models_df))]
}
# print(plt_models_df)
# print(myplot_radar(radar_inp_df=plt_models_df))
# print(myplot_radar(radar_inp_df=subset(plt_models_df,
# !(mdl_id %in% grep("random|MFO", plt_models_df$id, value=TRUE)))))
# Compute CI for <metric>SD
glb_models_df <- mutate(glb_models_df,
max.df = ifelse(max.nTuningRuns > 1, max.nTuningRuns - 1, NA),
min.sd2ci.scaler = ifelse(is.na(max.df), NA, qt(0.975, max.df)))
for (var in grep("SD", names(glb_models_df), value=TRUE)) {
# Does CI alredy exist ?
var_components <- unlist(strsplit(var, "SD"))
varActul <- paste0(var_components[1], var_components[2])
varUpper <- paste0(var_components[1], "Upper", var_components[2])
varLower <- paste0(var_components[1], "Lower", var_components[2])
if (varUpper %in% names(glb_models_df)) {
warning(varUpper, " already exists in glb_models_df")
# Assuming Lower also exists
next
}
print(sprintf("var:%s", var))
# CI is dependent on sample size in t distribution; df=n-1
glb_models_df[, varUpper] <- glb_models_df[, varActul] +
glb_models_df[, "min.sd2ci.scaler"] * glb_models_df[, var]
glb_models_df[, varLower] <- glb_models_df[, varActul] -
glb_models_df[, "min.sd2ci.scaler"] * glb_models_df[, var]
}## Warning: max.AccuracyUpper.fit already exists in glb_models_df## [1] "var:max.KappaSD.fit"# Plot metrics with CI
plt_models_df <- glb_models_df[, "id", FALSE]
pltCI_models_df <- glb_models_df[, "id", FALSE]
for (var in grep("Upper", names(glb_models_df), value=TRUE)) {
var_components <- unlist(strsplit(var, "Upper"))
col_name <- unlist(paste(var_components, collapse=""))
plt_models_df[, col_name] <- glb_models_df[, col_name]
for (name in paste0(var_components[1], c("Upper", "Lower"), var_components[2]))
pltCI_models_df[, name] <- glb_models_df[, name]
}
build_statsCI_data <- function(plt_models_df) {
mltd_models_df <- melt(plt_models_df, id.vars="id")
mltd_models_df$data <- sapply(1:nrow(mltd_models_df),
function(row_ix) tail(unlist(strsplit(as.character(
mltd_models_df[row_ix, "variable"]), "[.]")), 1))
mltd_models_df$label <- sapply(1:nrow(mltd_models_df),
function(row_ix) head(unlist(strsplit(as.character(
mltd_models_df[row_ix, "variable"]),
paste0(".", mltd_models_df[row_ix, "data"]))), 1))
#print(mltd_models_df)
return(mltd_models_df)
}
mltd_models_df <- build_statsCI_data(plt_models_df)
mltdCI_models_df <- melt(pltCI_models_df, id.vars="id")
for (row_ix in 1:nrow(mltdCI_models_df)) {
for (type in c("Upper", "Lower")) {
if (length(var_components <- unlist(strsplit(
as.character(mltdCI_models_df[row_ix, "variable"]), type))) > 1) {
#print(sprintf("row_ix:%d; type:%s; ", row_ix, type))
mltdCI_models_df[row_ix, "label"] <- var_components[1]
mltdCI_models_df[row_ix, "data"] <-
unlist(strsplit(var_components[2], "[.]"))[2]
mltdCI_models_df[row_ix, "type"] <- type
break
}
}
}
wideCI_models_df <- reshape(subset(mltdCI_models_df, select=-variable),
timevar="type",
idvar=setdiff(names(mltdCI_models_df), c("type", "value", "variable")),
direction="wide")
#print(wideCI_models_df)
mrgdCI_models_df <- merge(wideCI_models_df, mltd_models_df, all.x=TRUE)
#print(mrgdCI_models_df)
# Merge stats back in if CIs don't exist
goback_vars <- c()
for (var in unique(mltd_models_df$label)) {
for (type in unique(mltd_models_df$data)) {
var_type <- paste0(var, ".", type)
# if this data is already present, next
if (var_type %in% unique(paste(mltd_models_df$label, mltd_models_df$data,
sep=".")))
next
#print(sprintf("var_type:%s", var_type))
goback_vars <- c(goback_vars, var_type)
}
}
if (length(goback_vars) > 0) {
mltd_goback_df <- build_statsCI_data(glb_models_df[, c("id", goback_vars)])
mltd_models_df <- rbind(mltd_models_df, mltd_goback_df)
}
# mltd_models_df <- merge(mltd_models_df, glb_models_df[, c("id", "model_method")],
# all.x=TRUE)
png(paste0(glbOut$pfx, "models_bar.png"), width=480*3, height=480*2)
#print(gp <- myplot_bar(mltd_models_df, "id", "value", colorcol_name="model_method") +
print(gp <- myplot_bar(df=mltd_models_df, xcol_name="id", ycol_names="value") +
geom_errorbar(data=mrgdCI_models_df,
mapping=aes(x= id, ymax=value.Upper, ymin=value.Lower), width=0.5) +
facet_grid(label ~ data, scales="free") +
theme(axis.text.x = element_text(angle = 90,vjust = 0.5)))## Warning: Removed 4 rows containing missing values (geom_errorbar).dev.off()## quartz_off_screen
## 2print(gp)## Warning: Removed 4 rows containing missing values (geom_errorbar).
dsp_models_cols <- c("id",
glbMdlMetricsEval[glbMdlMetricsEval %in% names(glb_models_df)],
grep("opt.", names(glb_models_df), fixed = TRUE, value = TRUE))
# if (glb_is_classification && glb_is_binomial)
# dsp_models_cols <- c(dsp_models_cols, "opt.prob.threshold.OOB")
print(dsp_models_df <- orderBy(glbgetModelSelectFormula(), glb_models_df)[, dsp_models_cols])## id max.Accuracy.OOB max.AUCROCR.OOB
## 34 RFE.X#zv.pca#rcv#glmnet 0.5829596 0.5816510
## 22 All.X#nzv.spatialSign#rcv#glmnet 0.5820628 0.5787416
## 20 All.X#spatialSign#rcv#glmnet 0.5811659 0.5793099
## 38 RFE.X#nzv.spatialSign#rcv#glmnet 0.5811659 0.5785672
## 23 All.X.Inc#nzv.spatialSign#rcv#glmnet 0.5811659 0.5734523
## 36 RFE.X#spatialSign#rcv#glmnet 0.5802691 0.5794035
## 10 All.X#nzv#rcv#glmnet 0.5775785 0.5766539
## 26 RFE.X#nzv#rcv#glmnet 0.5775785 0.5762019
## 13 All.X#expoTrans#rcv#glmnet 0.5775785 0.5745421
## 29 RFE.X#expoTrans#rcv#glmnet 0.5775785 0.5745421
## 28 RFE.X#YeoJohnson#rcv#glmnet 0.5775785 0.5744824
## 12 All.X#YeoJohnson#rcv#glmnet 0.5775785 0.5744824
## 24 RFE.X##rcv#glmnet 0.5766816 0.5757966
## 25 RFE.X#zv#rcv#glmnet 0.5766816 0.5757966
## 27 RFE.X#BoxCox#rcv#glmnet 0.5766816 0.5757966
## 30 RFE.X#center#rcv#glmnet 0.5766816 0.5757966
## 31 RFE.X#scale#rcv#glmnet 0.5766816 0.5757966
## 32 RFE.X#center.scale#rcv#glmnet 0.5766816 0.5757966
## 33 RFE.X#range#rcv#glmnet 0.5766816 0.5757966
## 37 RFE.X#conditionalX#rcv#glmnet 0.5766816 0.5757966
## 6 Low.cor.X##rcv#glmnet 0.5766816 0.5757966
## 7 All.X##rcv#glmnet 0.5766816 0.5757966
## 9 All.X#zv#rcv#glmnet 0.5766816 0.5757966
## 11 All.X#BoxCox#rcv#glmnet 0.5766816 0.5757966
## 14 All.X#center#rcv#glmnet 0.5766816 0.5757966
## 15 All.X#scale#rcv#glmnet 0.5766816 0.5757966
## 16 All.X#center.scale#rcv#glmnet 0.5766816 0.5757966
## 17 All.X#range#rcv#glmnet 0.5766816 0.5757966
## 21 All.X#conditionalX#rcv#glmnet 0.5766816 0.5757966
## 18 All.X#zv.pca#rcv#glmnet 0.5748879 0.5798104
## 39 RFE.X#nzv.pca.spatialSign#rcv#glmnet 0.5713004 0.5742305
## 8 All.X##rcv#glm 0.5650224 0.5687249
## 35 RFE.X#ica#rcv#glmnet 0.5363229 0.5370894
## 19 All.X#ica#rcv#glmnet 0.5354260 0.5360626
## 5 Interact.High.cor.Y##rcv#glmnet 0.5345291 0.5242069
## 2 Random###myrandom_classfr 0.5300448 0.5181895
## 3 Max.cor.Y.rcv.1X1###glmnet 0.5300448 0.5102459
## 4 Max.cor.Y##rcv#rpart 0.5300448 0.5000646
## 1 MFO###myMFO_classfr 0.5300448 0.5000000
## max.AUCpROC.OOB max.Accuracy.fit opt.prob.threshold.fit
## 34 0.5792308 0.6459313 0.50
## 22 0.5572148 0.6478785 0.50
## 20 0.5570372 0.6492270 0.50
## 38 0.5597528 0.6462317 0.50
## 23 0.5556503 0.6440612 0.50
## 36 0.5570372 0.6490772 0.50
## 10 0.5550900 0.6475806 0.50
## 26 0.5531816 0.6466826 0.50
## 13 0.5559554 0.6463076 0.50
## 29 0.5559554 0.6462327 0.50
## 28 0.5559554 0.6466820 0.50
## 12 0.5559554 0.6465321 0.50
## 24 0.5588180 0.6475052 0.50
## 25 0.5588180 0.6475052 0.50
## 27 0.5588180 0.6475052 0.50
## 30 0.5588180 0.6475052 0.50
## 31 0.5588180 0.6475052 0.50
## 32 0.5588180 0.6475052 0.50
## 33 0.5588180 0.6475052 0.50
## 37 0.5588180 0.6475052 0.50
## 6 0.5588180 0.6471308 0.50
## 7 0.5588180 0.6471308 0.50
## 9 0.5588180 0.6471308 0.50
## 11 0.5588180 0.6471308 0.50
## 14 0.5588180 0.6471308 0.50
## 15 0.5588180 0.6471308 0.50
## 16 0.5588180 0.6471308 0.50
## 17 0.5588180 0.6471308 0.50
## 21 0.5588180 0.6471308 0.50
## 18 0.5717247 0.6421158 0.50
## 39 0.5545298 0.6480297 0.50
## 8 0.5487933 0.6254219 0.50
## 35 0.5121737 0.5603719 0.50
## 19 0.5133442 0.5607462 0.50
## 5 0.5218093 0.6250526 0.50
## 2 0.4836608 0.5299798 0.55
## 3 0.4999322 0.6240737 0.45
## 4 0.4999322 0.6227308 0.50
## 1 0.5000000 0.5299798 0.50
## opt.prob.threshold.OOB
## 34 0.50
## 22 0.60
## 20 0.60
## 38 0.60
## 23 0.60
## 36 0.60
## 10 0.60
## 26 0.55
## 13 0.55
## 29 0.55
## 28 0.55
## 12 0.55
## 24 0.55
## 25 0.55
## 27 0.55
## 30 0.55
## 31 0.55
## 32 0.55
## 33 0.55
## 37 0.55
## 6 0.55
## 7 0.55
## 9 0.55
## 11 0.55
## 14 0.55
## 15 0.55
## 16 0.55
## 17 0.55
## 21 0.55
## 18 0.50
## 39 0.55
## 8 0.65
## 35 0.55
## 19 0.55
## 5 0.70
## 2 0.55
## 3 0.65
## 4 0.65
## 1 0.50# print(myplot_radar(radar_inp_df = dsp_models_df))
print("Metrics used for model selection:"); print(glbgetModelSelectFormula())## [1] "Metrics used for model selection:"## ~-max.Accuracy.OOB - max.AUCROCR.OOB - max.AUCpROC.OOB - max.Accuracy.fit -
## opt.prob.threshold.OOB
## <environment: 0x7fa8afd836d8>print(sprintf("Best model id: %s", dsp_models_df[1, "id"]))## [1] "Best model id: RFE.X#zv.pca#rcv#glmnet"glb_get_predictions <- function(df, mdl_id, rsp_var, prob_threshold_def=NULL, verbose=FALSE) {
mdl <- glb_models_lst[[mdl_id]]
clmnNames <- mygetPredictIds(rsp_var, mdl_id)
predct_var_name <- clmnNames$value
predct_prob_var_name <- clmnNames$prob
predct_accurate_var_name <- clmnNames$is.acc
predct_error_var_name <- clmnNames$err
predct_erabs_var_name <- clmnNames$err.abs
if (glb_is_regression) {
df[, predct_var_name] <- predict(mdl, newdata=df, type="raw")
if (verbose) print(myplot_scatter(df, glb_rsp_var, predct_var_name) +
facet_wrap(reformulate(glbFeatsCategory), scales = "free") +
stat_smooth(method="glm"))
df[, predct_error_var_name] <- df[, predct_var_name] - df[, glb_rsp_var]
if (verbose) print(myplot_scatter(df, predct_var_name, predct_error_var_name) +
#facet_wrap(reformulate(glbFeatsCategory), scales = "free") +
stat_smooth(method="auto"))
if (verbose) print(myplot_scatter(df, glb_rsp_var, predct_error_var_name) +
#facet_wrap(reformulate(glbFeatsCategory), scales = "free") +
stat_smooth(method="glm"))
df[, predct_erabs_var_name] <- abs(df[, predct_error_var_name])
if (verbose) print(head(orderBy(reformulate(c("-", predct_erabs_var_name)), df)))
df[, predct_accurate_var_name] <- (df[, glb_rsp_var] == df[, predct_var_name])
}
if (glb_is_classification && glb_is_binomial) {
prob_threshold <- glb_models_df[glb_models_df$id == mdl_id,
"opt.prob.threshold.OOB"]
if (is.null(prob_threshold) || is.na(prob_threshold)) {
warning("Using default probability threshold: ", prob_threshold_def)
if (is.null(prob_threshold <- prob_threshold_def))
stop("Default probability threshold is NULL")
}
df[, predct_prob_var_name] <- predict(mdl, newdata = df, type = "prob")[, 2]
df[, predct_var_name] <-
factor(levels(df[, glb_rsp_var])[
(df[, predct_prob_var_name] >=
prob_threshold) * 1 + 1], levels(df[, glb_rsp_var]))
# if (verbose) print(myplot_scatter(df, glb_rsp_var, predct_var_name) +
# facet_wrap(reformulate(glbFeatsCategory), scales = "free") +
# stat_smooth(method="glm"))
df[, predct_error_var_name] <- df[, predct_var_name] != df[, glb_rsp_var]
# if (verbose) print(myplot_scatter(df, predct_var_name, predct_error_var_name) +
# #facet_wrap(reformulate(glbFeatsCategory), scales = "free") +
# stat_smooth(method="auto"))
# if (verbose) print(myplot_scatter(df, glb_rsp_var, predct_error_var_name) +
# #facet_wrap(reformulate(glbFeatsCategory), scales = "free") +
# stat_smooth(method="glm"))
# if prediction is a TP (true +ve), measure distance from 1.0
tp <- which((df[, predct_var_name] == df[, glb_rsp_var]) &
(df[, predct_var_name] == levels(df[, glb_rsp_var])[2]))
df[tp, predct_erabs_var_name] <- abs(1 - df[tp, predct_prob_var_name])
#rowIx <- which.max(df[tp, predct_erabs_var_name]); df[tp, c(glbFeatsId, glb_rsp_var, predct_var_name, predct_prob_var_name, predct_erabs_var_name)][rowIx, ]
# if prediction is a TN (true -ve), measure distance from 0.0
tn <- which((df[, predct_var_name] == df[, glb_rsp_var]) &
(df[, predct_var_name] == levels(df[, glb_rsp_var])[1]))
df[tn, predct_erabs_var_name] <- abs(0 - df[tn, predct_prob_var_name])
#rowIx <- which.max(df[tn, predct_erabs_var_name]); df[tn, c(glbFeatsId, glb_rsp_var, predct_var_name, predct_prob_var_name, predct_erabs_var_name)][rowIx, ]
# if prediction is a FP (flse +ve), measure distance from 0.0
fp <- which((df[, predct_var_name] != df[, glb_rsp_var]) &
(df[, predct_var_name] == levels(df[, glb_rsp_var])[2]))
df[fp, predct_erabs_var_name] <- abs(0 - df[fp, predct_prob_var_name])
#rowIx <- which.max(df[fp, predct_erabs_var_name]); df[fp, c(glbFeatsId, glb_rsp_var, predct_var_name, predct_prob_var_name, predct_erabs_var_name)][rowIx, ]
# if prediction is a FN (flse -ve), measure distance from 1.0
fn <- which((df[, predct_var_name] != df[, glb_rsp_var]) &
(df[, predct_var_name] == levels(df[, glb_rsp_var])[1]))
df[fn, predct_erabs_var_name] <- abs(1 - df[fn, predct_prob_var_name])
#rowIx <- which.max(df[fn, predct_erabs_var_name]); df[fn, c(glbFeatsId, glb_rsp_var, predct_var_name, predct_prob_var_name, predct_erabs_var_name)][rowIx, ]
if (verbose) print(head(orderBy(reformulate(c("-", predct_erabs_var_name)), df)))
df[, predct_accurate_var_name] <- (df[, glb_rsp_var] == df[, predct_var_name])
}
if (glb_is_classification && !glb_is_binomial) {
df[, predct_var_name] <- predict(mdl, newdata = df, type = "raw")
probCls <- predict(mdl, newdata = df, type = "prob")
df[, predct_prob_var_name] <- NA
for (cls in names(probCls)) {
mask <- (df[, predct_var_name] == cls)
df[mask, predct_prob_var_name] <- probCls[mask, cls]
}
if (verbose) print(myplot_histogram(df, predct_prob_var_name,
fill_col_name = predct_var_name))
if (verbose) print(myplot_histogram(df, predct_prob_var_name,
facet_frmla = paste0("~", glb_rsp_var)))
df[, predct_error_var_name] <- df[, predct_var_name] != df[, glb_rsp_var]
# if prediction is erroneous, measure predicted class prob from actual class prob
df[, predct_erabs_var_name] <- 0
for (cls in names(probCls)) {
mask <- (df[, glb_rsp_var] == cls) & (df[, predct_error_var_name])
df[mask, predct_erabs_var_name] <- probCls[mask, cls]
}
df[, predct_accurate_var_name] <- (df[, glb_rsp_var] == df[, predct_var_name])
}
return(df)
}
#stop(here"); glb2Sav(); glbObsAll <- savObsAll; glbObsTrn <- savObsTrn; glbObsFit <- savObsFit; glbObsOOB <- savObsOOB; sav_models_df <- glb_models_df; glb_models_df <- sav_models_df; glb_featsimp_df <- sav_featsimp_df
myget_category_stats <- function(obs_df, mdl_id, label) {
require(dplyr)
require(lazyeval)
predct_var_name <- mygetPredictIds(glb_rsp_var, mdl_id)$value
predct_error_var_name <- mygetPredictIds(glb_rsp_var, mdl_id)$err.abs
if (!predct_var_name %in% names(obs_df))
obs_df <- glb_get_predictions(obs_df, mdl_id, glb_rsp_var)
tmp_obs_df <- obs_df[, c(glbFeatsCategory, glb_rsp_var,
predct_var_name, predct_error_var_name)]
# tmp_obs_df <- obs_df %>%
# dplyr::select_(glbFeatsCategory, glb_rsp_var, predct_var_name, predct_error_var_name)
#dplyr::rename(startprice.log10.predict.RFE.X.glmnet.err=error_abs_OOB)
names(tmp_obs_df)[length(names(tmp_obs_df))] <- paste0("err.abs.", label)
ret_ctgry_df <- tmp_obs_df %>%
dplyr::group_by_(glbFeatsCategory) %>%
dplyr::summarise_(#interp(~sum(abs(var)), var=as.name(glb_rsp_var)),
interp(~sum(var), var=as.name(paste0("err.abs.", label))),
interp(~mean(var), var=as.name(paste0("err.abs.", label))),
interp(~n()))
names(ret_ctgry_df) <- c(glbFeatsCategory,
#paste0(glb_rsp_var, ".abs.", label, ".sum"),
paste0("err.abs.", label, ".sum"),
paste0("err.abs.", label, ".mean"),
paste0(".n.", label))
ret_ctgry_df <- dplyr::ungroup(ret_ctgry_df)
#colSums(ret_ctgry_df[, -grep(glbFeatsCategory, names(ret_ctgry_df))])
return(ret_ctgry_df)
}
#print(colSums((ctgry_df <- myget_category_stats(obs_df=glbObsFit, mdl_id="", label="fit"))[, -grep(glbFeatsCategory, names(ctgry_df))]))
if (!is.null(glb_mdl_ensemble)) {
fit.models_2_chunk_df <- myadd_chunk(fit.models_2_chunk_df,
paste0("fit.models_2_", mdl_id_pfx), major.inc = TRUE,
label.minor = "ensemble")
mdl_id_pfx <- "Ensemble"
if (#(glb_is_regression) |
((glb_is_classification) & (!glb_is_binomial)))
stop("Ensemble models not implemented yet for multinomial classification")
mygetEnsembleAutoMdlIds <- function() {
tmp_models_df <- orderBy(glbgetModelSelectFormula(), glb_models_df)
row.names(tmp_models_df) <- tmp_models_df$id
mdl_threshold_pos <-
min(which(grepl("MFO|Random|Baseline", tmp_models_df$id))) - 1
mdlIds <- tmp_models_df$id[1:mdl_threshold_pos]
return(mdlIds[!grepl("Ensemble", mdlIds)])
}
if (glb_mdl_ensemble == "auto") {
glb_mdl_ensemble <- mygetEnsembleAutoMdlIds()
mdl_id_pfx <- paste0(mdl_id_pfx, ".auto")
} else if (grepl("^%<d-%", glb_mdl_ensemble)) {
glb_mdl_ensemble <- eval(parse(text =
str_trim(unlist(strsplit(glb_mdl_ensemble, "%<d-%"))[2])))
}
for (mdl_id in glb_mdl_ensemble) {
if (!(mdl_id %in% names(glb_models_lst))) {
warning("Model ", mdl_id, " in glb_model_ensemble not found !")
next
}
glbObsFit <- glb_get_predictions(df = glbObsFit, mdl_id, glb_rsp_var)
glbObsOOB <- glb_get_predictions(df = glbObsOOB, mdl_id, glb_rsp_var)
}
#mdl_id_pfx <- "Ensemble.RFE"; mdlId <- paste0(mdl_id_pfx, ".glmnet")
#glb_mdl_ensemble <- gsub(mygetPredictIds$value, "", grep("RFE\\.X\\.(?!Interact)", row.names(glb_featsimp_df), perl = TRUE, value = TRUE), fixed = TRUE)
#varImp(glb_models_lst[[mdlId]])
#cor_df <- data.frame(cor=cor(glbObsFit[, glb_rsp_var], glbObsFit[, paste(mygetPredictIds$value, glb_mdl_ensemble)], use="pairwise.complete.obs"))
#glbObsFit <- glb_get_predictions(df=glbObsFit, "Ensemble.glmnet", glb_rsp_var);print(colSums((ctgry_df <- myget_category_stats(obs_df=glbObsFit, mdl_id="Ensemble.glmnet", label="fit"))[, -grep(glbFeatsCategory, names(ctgry_df))]))
### bid0_sp
# Better than MFO; models.n=28; min.RMSE.fit=0.0521233; err.abs.fit.sum=7.3631895
# old: Top x from auto; models.n= 5; min.RMSE.fit=0.06311047; err.abs.fit.sum=9.5937080
# RFE only ; models.n=16; min.RMSE.fit=0.05148588; err.abs.fit.sum=7.2875091
# RFE subset only ;models.n= 5; min.RMSE.fit=0.06040702; err.abs.fit.sum=9.059088
# RFE subset only ;models.n= 9; min.RMSE.fit=0.05933167; err.abs.fit.sum=8.7421288
# RFE subset only ;models.n=15; min.RMSE.fit=0.0584607; err.abs.fit.sum=8.5902066
# RFE subset only ;models.n=17; min.RMSE.fit=0.05496899; err.abs.fit.sum=8.0170431
# RFE subset only ;models.n=18; min.RMSE.fit=0.05441577; err.abs.fit.sum=7.837223
# RFE subset only ;models.n=16; min.RMSE.fit=0.05441577; err.abs.fit.sum=7.837223
### bid0_sp
### bid1_sp
# "auto"; err.abs.fit.sum=76.699774; min.RMSE.fit=0.2186429
# "RFE.X.*"; err.abs.fit.sum=; min.RMSE.fit=0.221114
### bid1_sp
indepVar <- paste(mygetPredictIds(glb_rsp_var)$value, glb_mdl_ensemble, sep = "")
if (glb_is_classification)
indepVar <- paste(indepVar, ".prob", sep = "")
# Some models in glb_mdl_ensemble might not be fitted e.g. RFE.X.Interact
indepVar <- intersect(indepVar, names(glbObsFit))
# indepVar <- grep(mygetPredictIds(glb_rsp_var)$value, names(glbObsFit), fixed=TRUE, value=TRUE)
# if (glb_is_regression)
# indepVar <- indepVar[!grepl("(err\\.abs|accurate)$", indepVar)]
# if (glb_is_classification && glb_is_binomial)
# indepVar <- grep("prob$", indepVar, value=TRUE) else
# indepVar <- indepVar[!grepl("err$", indepVar)]
#rfe_fit_ens_results <- myrun_rfe(glbObsFit, indepVar)
for (method in c("glm", "glmnet")) {
for (trainControlMethod in
c("boot", "boot632", "cv", "repeatedcv"
#, "LOOCV" # tuneLength * nrow(fitDF)
, "LGOCV", "adaptive_cv"
#, "adaptive_boot" #error: adaptive$min should be less than 3
#, "adaptive_LGOCV" #error: adaptive$min should be less than 3
)) {
#sav_models_df <- glb_models_df; all.equal(sav_models_df, glb_models_df)
#glb_models_df <- sav_models_df; print(glb_models_df$id)
if ((method == "glm") && (trainControlMethod != "repeatedcv"))
# glm used only to identify outliers
next
ret_lst <- myfit_mdl(
mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = paste0(mdl_id_pfx, ".", trainControlMethod),
type = glb_model_type, tune.df = NULL,
trainControl.method = trainControlMethod,
trainControl.number = glb_rcv_n_folds,
trainControl.repeats = glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method = method)),
indepVar = indepVar, rsp_var = glb_rsp_var,
fit_df = glbObsFit, OOB_df = glbObsOOB)
}
}
dsp_models_df <- glbgetDisplayModelsDf()
}
if (is.null(glbMdlSelId))
glbMdlSelId <- dsp_models_df[1, "id"] else
print(sprintf("User specified selection: %s", glbMdlSelId))
# knitr crashes sometimes with plot.glmnet
if (!(is.null(knitr::opts_current$get(name = 'label'))) &&
(myparseMdlId(glbMdlSelId)$alg == "glmnet"))
print(glb_sel_mdl <- glb_models_lst[[glbMdlSelId]]) else
myprint_mdl(glb_sel_mdl <- glb_models_lst[[glbMdlSelId]])## glmnet
##
## Pre-processing: principal component signal extraction (243),
## centered (243), scaled (243), remove (1)
## Resampling: Cross-Validated (3 fold, repeated 3 times)
## Summary of sample sizes: 2969, 2969, 2968, 2968, 2970, 2968, ...
## Resampling results across tuning parameters:
##
## alpha lambda Accuracy Kappa
## 0.100 8.883582e-05 0.6380735 0.2733807
## 0.100 4.123393e-04 0.6380735 0.2733807
## 0.100 1.913910e-03 0.6381484 0.2734989
## 0.100 8.883582e-03 0.6394212 0.2756833
## 0.100 4.123393e-02 0.6436874 0.2823364
## 0.325 8.883582e-05 0.6382981 0.2738119
## 0.325 4.123393e-04 0.6382981 0.2738119
## 0.325 1.913910e-03 0.6386718 0.2744242
## 0.325 8.883582e-03 0.6426389 0.2816353
## 0.325 4.123393e-02 0.6412897 0.2761086
## 0.550 8.883582e-05 0.6385972 0.2744123
## 0.550 4.123393e-04 0.6385972 0.2744123
## 0.550 1.913910e-03 0.6395700 0.2761095
## 0.550 8.883582e-03 0.6440610 0.2840948
## 0.550 4.123393e-02 0.6323820 0.2562114
## 0.775 8.883582e-05 0.6385223 0.2742817
## 0.775 4.123393e-04 0.6385224 0.2743094
## 0.775 1.913910e-03 0.6391210 0.2750959
## 0.775 8.883582e-03 0.6459313 0.2875399
## 0.775 4.123393e-02 0.6232505 0.2355461
## 1.000 8.883582e-05 0.6387470 0.2747400
## 1.000 4.123393e-04 0.6382230 0.2736425
## 1.000 1.913910e-03 0.6410669 0.2788996
## 1.000 8.883582e-03 0.6458553 0.2868669
## 1.000 4.123393e-02 0.6177872 0.2213701
##
## Accuracy was used to select the optimal model using the largest value.
## The final values used for the model were alpha = 0.775 and lambda
## = 0.008883582.# From here to save(), this should all be in one function
# these are executed in the same seq twice more:
# fit.data.training & predict.data.new chunks
print(sprintf("%s fit prediction diagnostics:", glbMdlSelId))## [1] "RFE.X#zv.pca#rcv#glmnet fit prediction diagnostics:"glbObsFit <- glb_get_predictions(df = glbObsFit, mdl_id = glbMdlSelId,
rsp_var = glb_rsp_var)
print(sprintf("%s OOB prediction diagnostics:", glbMdlSelId))## [1] "RFE.X#zv.pca#rcv#glmnet OOB prediction diagnostics:"glbObsOOB <- glb_get_predictions(df = glbObsOOB, mdl_id = glbMdlSelId,
rsp_var = glb_rsp_var)
print(glb_featsimp_df <- myget_feats_importance(mdl = glb_sel_mdl, featsimp_df = NULL))## RFE.X.zv.pca.rcv.glmnet.imp imp
## PC8 1.000000e+02 1.000000e+02
## PC5 7.995721e+01 7.995721e+01
## PC10 6.660214e+01 6.660214e+01
## PC15 6.321726e+01 6.321726e+01
## PC9 5.584978e+01 5.584978e+01
## PC26 5.050796e+01 5.050796e+01
## PC3 4.474537e+01 4.474537e+01
## PC14 4.398423e+01 4.398423e+01
## PC57 3.105066e+01 3.105066e+01
## PC68 2.684775e+01 2.684775e+01
## PC30 2.569378e+01 2.569378e+01
## PC88 2.477208e+01 2.477208e+01
## PC128 2.424400e+01 2.424400e+01
## PC25 2.371905e+01 2.371905e+01
## PC19 2.341047e+01 2.341047e+01
## PC124 2.213194e+01 2.213194e+01
## PC52 2.128227e+01 2.128227e+01
## PC16 2.054784e+01 2.054784e+01
## PC142 1.751430e+01 1.751430e+01
## PC114 1.650518e+01 1.650518e+01
## PC40 1.496730e+01 1.496730e+01
## PC23 1.488607e+01 1.488607e+01
## PC22 1.452383e+01 1.452383e+01
## PC100 1.448174e+01 1.448174e+01
## PC138 1.321459e+01 1.321459e+01
## PC126 1.295801e+01 1.295801e+01
## PC45 1.271547e+01 1.271547e+01
## PC101 1.257034e+01 1.257034e+01
## PC2 1.110115e+01 1.110115e+01
## PC11 1.060646e+01 1.060646e+01
## PC77 8.954286e+00 8.954286e+00
## PC6 8.539962e+00 8.539962e+00
## PC17 8.207663e+00 8.207663e+00
## PC70 7.297321e+00 7.297321e+00
## PC47 6.727046e+00 6.727046e+00
## PC106 6.709574e+00 6.709574e+00
## PC27 6.583676e+00 6.583676e+00
## PC108 6.523922e+00 6.523922e+00
## PC20 6.471282e+00 6.471282e+00
## PC95 6.273683e+00 6.273683e+00
## PC92 6.234515e+00 6.234515e+00
## PC111 5.951208e+00 5.951208e+00
## PC87 5.856015e+00 5.856015e+00
## PC117 5.778034e+00 5.778034e+00
## PC144 5.666493e+00 5.666493e+00
## PC54 5.627894e+00 5.627894e+00
## PC74 5.526187e+00 5.526187e+00
## PC140 5.171006e+00 5.171006e+00
## PC85 5.090498e+00 5.090498e+00
## PC118 4.902623e+00 4.902623e+00
## PC109 4.849634e+00 4.849634e+00
## PC44 4.657604e+00 4.657604e+00
## PC4 4.495763e+00 4.495763e+00
## PC35 4.402217e+00 4.402217e+00
## PC84 3.861871e+00 3.861871e+00
## PC102 3.345513e+00 3.345513e+00
## PC49 3.217595e+00 3.217595e+00
## PC113 2.752156e+00 2.752156e+00
## PC103 2.182427e+00 2.182427e+00
## PC130 2.081387e+00 2.081387e+00
## PC51 1.903814e+00 1.903814e+00
## PC33 1.735816e+00 1.735816e+00
## PC1 1.487771e+00 1.487771e+00
## PC75 1.133188e+00 1.133188e+00
## PC123 4.889254e-01 4.889254e-01
## PC99 3.645912e-01 3.645912e-01
## PC46 2.206416e-01 2.206416e-01
## PC12 1.279928e-01 1.279928e-01
## PC133 8.569951e-02 8.569951e-02
## PC137 7.388924e-02 7.388924e-02
## PC37 2.029163e-02 2.029163e-02
## PC120 7.951659e-03 7.951659e-03
## PC7 0.000000e+00 0.000000e+00
## PC13 0.000000e+00 0.000000e+00
## PC18 0.000000e+00 0.000000e+00
## PC21 0.000000e+00 0.000000e+00
## PC24 0.000000e+00 0.000000e+00
## PC28 0.000000e+00 0.000000e+00
## PC29 0.000000e+00 0.000000e+00
## PC31 0.000000e+00 0.000000e+00
## PC32 0.000000e+00 0.000000e+00
## PC34 0.000000e+00 0.000000e+00
## PC36 0.000000e+00 0.000000e+00
## PC38 0.000000e+00 0.000000e+00
## PC39 0.000000e+00 0.000000e+00
## PC41 0.000000e+00 0.000000e+00
## PC42 0.000000e+00 0.000000e+00
## PC43 0.000000e+00 0.000000e+00
## PC48 0.000000e+00 0.000000e+00
## PC50 0.000000e+00 0.000000e+00
## PC53 0.000000e+00 0.000000e+00
## PC55 0.000000e+00 0.000000e+00
## PC56 0.000000e+00 0.000000e+00
## PC58 0.000000e+00 0.000000e+00
## PC59 0.000000e+00 0.000000e+00
## PC60 0.000000e+00 0.000000e+00
## PC61 0.000000e+00 0.000000e+00
## PC62 0.000000e+00 0.000000e+00
## PC63 0.000000e+00 0.000000e+00
## PC64 0.000000e+00 0.000000e+00
## PC65 0.000000e+00 0.000000e+00
## PC66 0.000000e+00 0.000000e+00
## PC67 0.000000e+00 0.000000e+00
## PC69 0.000000e+00 0.000000e+00
## PC71 0.000000e+00 0.000000e+00
## PC72 0.000000e+00 0.000000e+00
## PC73 0.000000e+00 0.000000e+00
## PC76 0.000000e+00 0.000000e+00
## PC78 0.000000e+00 0.000000e+00
## PC79 0.000000e+00 0.000000e+00
## PC80 0.000000e+00 0.000000e+00
## PC81 0.000000e+00 0.000000e+00
## PC82 0.000000e+00 0.000000e+00
## PC83 0.000000e+00 0.000000e+00
## PC86 0.000000e+00 0.000000e+00
## PC89 0.000000e+00 0.000000e+00
## PC90 0.000000e+00 0.000000e+00
## PC91 0.000000e+00 0.000000e+00
## PC93 0.000000e+00 0.000000e+00
## PC94 0.000000e+00 0.000000e+00
## PC96 0.000000e+00 0.000000e+00
## PC97 0.000000e+00 0.000000e+00
## PC98 0.000000e+00 0.000000e+00
## PC104 0.000000e+00 0.000000e+00
## PC105 0.000000e+00 0.000000e+00
## PC107 0.000000e+00 0.000000e+00
## PC110 0.000000e+00 0.000000e+00
## PC112 0.000000e+00 0.000000e+00
## PC115 0.000000e+00 0.000000e+00
## PC116 0.000000e+00 0.000000e+00
## PC119 0.000000e+00 0.000000e+00
## PC121 0.000000e+00 0.000000e+00
## PC122 0.000000e+00 0.000000e+00
## PC125 0.000000e+00 0.000000e+00
## PC127 0.000000e+00 0.000000e+00
## PC129 0.000000e+00 0.000000e+00
## PC131 0.000000e+00 0.000000e+00
## PC132 0.000000e+00 0.000000e+00
## PC134 0.000000e+00 0.000000e+00
## PC135 0.000000e+00 0.000000e+00
## PC136 0.000000e+00 0.000000e+00
## PC139 0.000000e+00 0.000000e+00
## PC141 0.000000e+00 0.000000e+00
## PC143 0.000000e+00 0.000000e+00
## PC145 0.000000e+00 0.000000e+00
## PC146 0.000000e+00 0.000000e+00#mdl_id <-"RFE.X.glmnet"; glb_featsimp_df <- myget_feats_importance(glb_models_lst[[mdl_id]], glb_featsimp_df); glb_featsimp_df[, paste0(mdl_id, ".imp")] <- glb_featsimp_df$imp; print(glb_featsimp_df)
#print(head(sbst_featsimp_df <- subset(glb_featsimp_df, is.na(RFE.X.glmnet.imp) | (abs(RFE.X.YeoJohnson.glmnet.imp - RFE.X.glmnet.imp) > 0.0001), select=-imp)))
#print(orderBy(~ -cor.y.abs, subset(glb_feats_df, id %in% c(row.names(sbst_featsimp_df), "startprice.dcm1.is9", "D.weight.post.stop.sum"))))
# Used again in fit.data.training & predict.data.new chunks
glb_analytics_diag_plots <- function(obs_df, mdl_id, prob_threshold=NULL) {
if (!is.null(featsimp_df <- glb_featsimp_df)) {
featsimp_df$feat <- gsub("`(.*?)`", "\\1", row.names(featsimp_df))
featsimp_df$feat.interact <- gsub("(.*?):(.*)", "\\2", featsimp_df$feat)
featsimp_df$feat <- gsub("(.*?):(.*)", "\\1", featsimp_df$feat)
featsimp_df$feat.interact <-
ifelse(featsimp_df$feat.interact == featsimp_df$feat,
NA, featsimp_df$feat.interact)
featsimp_df$feat <-
gsub("(.*?)\\.fctr(.*)", "\\1\\.fctr", featsimp_df$feat)
featsimp_df$feat.interact <-
gsub("(.*?)\\.fctr(.*)", "\\1\\.fctr", featsimp_df$feat.interact)
featsimp_df <- orderBy(~ -imp.max,
summaryBy(imp ~ feat + feat.interact, data=featsimp_df,
FUN=max))
#rex_str=":(.*)"; txt_vctr=tail(featsimp_df$feat); ret_lst <- regexec(rex_str, txt_vctr); ret_lst <- regmatches(txt_vctr, ret_lst); ret_vctr <- sapply(1:length(ret_lst), function(pos_ix) ifelse(length(ret_lst[[pos_ix]]) > 0, ret_lst[[pos_ix]], "")); print(ret_vctr <- ret_vctr[ret_vctr != ""])
featsimp_df <- subset(featsimp_df, !is.na(imp.max))
if (nrow(featsimp_df) > 5) {
warning("Limiting important feature scatter plots to 5 out of ",
nrow(featsimp_df))
featsimp_df <- head(featsimp_df, 5)
}
# if (!all(is.na(featsimp_df$feat.interact)))
# stop("not implemented yet")
rsp_var_out <- mygetPredictIds(glb_rsp_var, mdl_id)$value
if (grepl("pca", myparseMdlId(mdl_id)$preProc, fixed = TRUE)) {
indepVar <- unlist(strsplit(glb_models_df[glb_models_df$id == mdl_id, "feats"], ","))
vectorizedObsMtx <- myget_vectorized_obs_df(obs_df, rsp_var = glb_rsp_var,
indep_vars = indepVar)
if (!inherits(vectorizedObsMtx, "matrix")) {
vectorizedObsMtx[, glb_rsp_var] <- as.numeric(vectorizedObsMtx[, glb_rsp_var])
vectorizedObsMtx <- as.matrix(vectorizedObsMtx)
}
rotationMtx <- glb_models_lst[[mdl_id]]$preProcess$rotation
pcaobs_df <- as.data.frame(vectorizedObsMtx[, dimnames(rotationMtx)[[1]]] %*%
rotationMtx)
# pcaobs_df[, glb_rsp_var] <- obs_df[, glb_rsp_var]
# pcaobs_df[, rsp_var_out] <- obs_df[, rsp_var_out]
# pcaobs_df[, glbFeatsId] <- obs_df[, glbFeatsId]
obs_df <- cbind(obs_df, pcaobs_df)
}
for (var in featsimp_df$feat) {
plot_df <- melt(obs_df, id.vars = var,
measure.vars = c(glb_rsp_var, rsp_var_out))
print(myplot_scatter(plot_df, var, "value", colorcol_name = "variable",
facet_colcol_name = "variable", jitter = TRUE) +
guides(color = FALSE))
}
}
if (glb_is_regression) {
if (is.null(featsimp_df) || (nrow(featsimp_df) == 0))
warning("No important features in glb_fin_mdl") else
print(myplot_prediction_regression(df=obs_df,
feat_x=ifelse(nrow(featsimp_df) > 1, featsimp_df$feat[2],
".rownames"),
feat_y=featsimp_df$feat[1],
rsp_var=glb_rsp_var, rsp_var_out=rsp_var_out,
id_vars=glbFeatsId)
# + facet_wrap(reformulate(featsimp_df$feat[2])) # if [1 or 2] is a factor
# + geom_point(aes_string(color="<col_name>.fctr")) # to color the plot
)
}
if (glb_is_classification) {
require(lazyeval)
if (is.null(featsimp_df) || (nrow(featsimp_df) == 0))
warning("No features in selected model are statistically important") else
print(myplot_prediction_classification(df = obs_df,
feat_x = ifelse(nrow(featsimp_df) > 1,
featsimp_df$feat[2], ".rownames"),
feat_y = featsimp_df$feat[1],
rsp_var = glb_rsp_var,
rsp_var_out = rsp_var_out,
id_vars = glbFeatsId,
prob_threshold = prob_threshold))
}
}
if (glb_is_classification && glb_is_binomial)
glb_analytics_diag_plots(obs_df = glbObsOOB, mdl_id = glbMdlSelId,
prob_threshold = glb_models_df[glb_models_df$id == glbMdlSelId,
"opt.prob.threshold.OOB"]) else
glb_analytics_diag_plots(obs_df = glbObsOOB, mdl_id = glbMdlSelId) ## Warning in glb_analytics_diag_plots(obs_df = glbObsOOB, mdl_id =
## glbMdlSelId, : Limiting important feature scatter plots to 5 out of 146
## Loading required package: lazyeval
## [1] "Min/Max Boundaries: "
## USER_ID Party.fctr Party.fctr.RFE.X.zv.pca.rcv.glmnet.prob
## 1 3908 D 0.13999479
## 2 4498 D 0.08080308
## 3 48 D 0.59038150
## 4 5858 D 0.70569556
## Party.fctr.RFE.X.zv.pca.rcv.glmnet
## 1 D
## 2 D
## 3 R
## 4 R
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.err
## 1 FALSE
## 2 FALSE
## 3 TRUE
## 4 TRUE
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.err.abs
## 1 0.13999479
## 2 0.08080308
## 3 0.59038150
## 4 0.70569556
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.is.acc
## 1 TRUE
## 2 TRUE
## 3 FALSE
## 4 FALSE
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.accurate
## 1 TRUE
## 2 TRUE
## 3 FALSE
## 4 FALSE
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.error .label
## 1 0.0000000 3908
## 2 0.0000000 4498
## 3 0.0903815 48
## 4 0.2056956 5858
## [1] "Inaccurate: "
## USER_ID Party.fctr Party.fctr.RFE.X.zv.pca.rcv.glmnet.prob
## 1 3895 R 0.06556290
## 2 1515 R 0.07338278
## 3 468 R 0.07477716
## 4 1653 R 0.08336510
## 5 1307 R 0.08611118
## 6 3878 R 0.09605418
## Party.fctr.RFE.X.zv.pca.rcv.glmnet
## 1 D
## 2 D
## 3 D
## 4 D
## 5 D
## 6 D
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.err
## 1 TRUE
## 2 TRUE
## 3 TRUE
## 4 TRUE
## 5 TRUE
## 6 TRUE
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.err.abs
## 1 0.9344371
## 2 0.9266172
## 3 0.9252228
## 4 0.9166349
## 5 0.9138888
## 6 0.9039458
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.is.acc
## 1 FALSE
## 2 FALSE
## 3 FALSE
## 4 FALSE
## 5 FALSE
## 6 FALSE
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.accurate
## 1 FALSE
## 2 FALSE
## 3 FALSE
## 4 FALSE
## 5 FALSE
## 6 FALSE
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.error
## 1 -0.4344371
## 2 -0.4266172
## 3 -0.4252228
## 4 -0.4166349
## 5 -0.4138888
## 6 -0.4039458
## USER_ID Party.fctr Party.fctr.RFE.X.zv.pca.rcv.glmnet.prob
## 104 277 R 0.3674392
## 211 5162 R 0.4757984
## 268 4298 D 0.5121844
## 385 6880 D 0.6100009
## 429 477 D 0.6963740
## 432 1765 D 0.7015358
## Party.fctr.RFE.X.zv.pca.rcv.glmnet
## 104 D
## 211 D
## 268 R
## 385 R
## 429 R
## 432 R
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.err
## 104 TRUE
## 211 TRUE
## 268 TRUE
## 385 TRUE
## 429 TRUE
## 432 TRUE
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.err.abs
## 104 0.6325608
## 211 0.5242016
## 268 0.5121844
## 385 0.6100009
## 429 0.6963740
## 432 0.7015358
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.is.acc
## 104 FALSE
## 211 FALSE
## 268 FALSE
## 385 FALSE
## 429 FALSE
## 432 FALSE
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.accurate
## 104 FALSE
## 211 FALSE
## 268 FALSE
## 385 FALSE
## 429 FALSE
## 432 FALSE
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.error
## 104 -0.13256075
## 211 -0.02420158
## 268 0.01218441
## 385 0.11000095
## 429 0.19637398
## 432 0.20153577
## USER_ID Party.fctr Party.fctr.RFE.X.zv.pca.rcv.glmnet.prob
## 460 892 D 0.8034368
## 461 4859 D 0.8062765
## 462 3978 D 0.8089729
## 463 1309 D 0.8106673
## 464 217 D 0.8319108
## 465 4276 D 0.8587556
## Party.fctr.RFE.X.zv.pca.rcv.glmnet
## 460 R
## 461 R
## 462 R
## 463 R
## 464 R
## 465 R
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.err
## 460 TRUE
## 461 TRUE
## 462 TRUE
## 463 TRUE
## 464 TRUE
## 465 TRUE
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.err.abs
## 460 0.8034368
## 461 0.8062765
## 462 0.8089729
## 463 0.8106673
## 464 0.8319108
## 465 0.8587556
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.is.acc
## 460 FALSE
## 461 FALSE
## 462 FALSE
## 463 FALSE
## 464 FALSE
## 465 FALSE
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.accurate
## 460 FALSE
## 461 FALSE
## 462 FALSE
## 463 FALSE
## 464 FALSE
## 465 FALSE
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.error
## 460 0.3034368
## 461 0.3062765
## 462 0.3089729
## 463 0.3106673
## 464 0.3319108
## 465 0.3587556
if (!is.null(glbFeatsCategory)) {
glbLvlCategory <- merge(glbLvlCategory,
myget_category_stats(obs_df = glbObsFit, mdl_id = glbMdlSelId,
label = "fit"),
by = glbFeatsCategory, all = TRUE)
row.names(glbLvlCategory) <- glbLvlCategory[, glbFeatsCategory]
glbLvlCategory <- merge(glbLvlCategory,
myget_category_stats(obs_df = glbObsOOB, mdl_id = glbMdlSelId,
label="OOB"),
#by=glbFeatsCategory, all=TRUE) glb_ctgry-df already contains .n.OOB ?
all = TRUE)
row.names(glbLvlCategory) <- glbLvlCategory[, glbFeatsCategory]
if (any(grepl("OOB", glbMdlMetricsEval)))
print(orderBy(~-err.abs.OOB.mean, glbLvlCategory)) else
print(orderBy(~-err.abs.fit.mean, glbLvlCategory))
print(colSums(glbLvlCategory[, -grep(glbFeatsCategory, names(glbLvlCategory))]))
}## Q109244.fctr .n.OOB .n.Fit .n.Tst .freqRatio.Fit .freqRatio.OOB
## No No 498 1961 622 0.4403773 0.4466368
## NA NA 438 1746 547 0.3920952 0.3928251
## Yes Yes 179 746 223 0.1675275 0.1605381
## .freqRatio.Tst err.abs.fit.sum err.abs.fit.mean .n.fit err.abs.OOB.sum
## No 0.4468391 883.9618 0.4507709 1961 239.90734
## NA 0.3929598 821.7109 0.4706248 1746 209.08425
## Yes 0.1602011 200.4801 0.2687400 746 85.28436
## err.abs.OOB.mean
## No 0.4817416
## NA 0.4773613
## Yes 0.4764489
## .n.OOB .n.Fit .n.Tst .freqRatio.Fit
## 1115.000000 4453.000000 1392.000000 1.000000
## .freqRatio.OOB .freqRatio.Tst err.abs.fit.sum err.abs.fit.mean
## 1.000000 1.000000 1906.152762 1.190136
## .n.fit err.abs.OOB.sum err.abs.OOB.mean
## 4453.000000 534.275955 1.435552write.csv(glbObsOOB[, c(glbFeatsId,
grep(glb_rsp_var, names(glbObsOOB), fixed=TRUE, value=TRUE))],
paste0(gsub(".", "_", paste0(glbOut$pfx, glbMdlSelId), fixed=TRUE),
"_OOBobs.csv"), row.names=FALSE)
fit.models_2_chunk_df <-
myadd_chunk(NULL, "fit.models_2_bgn", label.minor = "teardown")## label step_major step_minor label_minor bgn end elapsed
## 1 fit.models_2_bgn 1 0 teardown 334.101 NA NAglb_chunks_df <- myadd_chunk(glb_chunks_df, "fit.models", major.inc=FALSE)## label step_major step_minor label_minor bgn end elapsed
## 2 fit.models 1 1 1 319.665 334.11 14.445
## 3 fit.models 1 2 2 334.111 NA NA# if (sum(is.na(glbObsAll$D.P.http)) > 0)
# stop("fit.models_3: Why is this happening ?")
#stop(here"); glb2Sav()
sync_glb_obs_df <- function() {
# Merge or cbind ?
for (col in setdiff(names(glbObsFit), names(glbObsTrn)))
glbObsTrn[glbObsTrn$.lcn == "Fit", col] <<- glbObsFit[, col]
for (col in setdiff(names(glbObsFit), names(glbObsAll)))
glbObsAll[glbObsAll$.lcn == "Fit", col] <<- glbObsFit[, col]
if (all(is.na(glbObsNew[, glb_rsp_var])))
for (col in setdiff(names(glbObsOOB), names(glbObsTrn)))
glbObsTrn[glbObsTrn$.lcn == "OOB", col] <<- glbObsOOB[, col]
for (col in setdiff(names(glbObsOOB), names(glbObsAll)))
glbObsAll[glbObsAll$.lcn == "OOB", col] <<- glbObsOOB[, col]
}
sync_glb_obs_df()
print(setdiff(names(glbObsNew), names(glbObsAll)))## character(0)replay.petrisim(pn = glb_analytics_pn,
replay.trans = (glb_analytics_avl_objs <- c(glb_analytics_avl_objs,
"model.selected")), flip_coord = TRUE)## time trans "bgn " "fit.data.training.all " "predict.data.new " "end "
## 0.0000 multiple enabled transitions: data.training.all data.new model.selected firing: model.selected
## 1.0000 3 2 1 0 0
glb_chunks_df <- myadd_chunk(glb_chunks_df, "fit.data.training", major.inc = TRUE)## label step_major step_minor label_minor bgn end
## 3 fit.models 1 2 2 334.111 339.957
## 4 fit.data.training 2 0 0 339.957 NA
## elapsed
## 3 5.846
## 4 NAStep 2.0: fit data training
#load(paste0(glb_inp_pfx, "dsk.RData"))
if (!is.null(glbMdlFinId) && (glbMdlFinId %in% names(glb_models_lst))) {
warning("Final model same as user selected model")
glb_fin_mdl <- glb_models_lst[[glbMdlFinId]]
} else
# if (nrow(glbObsFit) + length(glbObsFitOutliers) == nrow(glbObsTrn))
if (!all(is.na(glbObsNew[, glb_rsp_var]))) {
warning("Final model same as glbMdlSelId")
glbMdlFinId <- paste0("Final.", glbMdlSelId)
glb_fin_mdl <- glb_sel_mdl
glb_models_lst[[glbMdlFinId]] <- glb_fin_mdl
mdlDf <- glb_models_df[glb_models_df$id == glbMdlSelId, ]
mdlDf$id <- glbMdlFinId
glb_models_df <- rbind(glb_models_df, mdlDf)
} else {
if (myparseMdlId(glbMdlSelId)$family == "RFE.X") {
indepVar <- mygetIndepVar(glb_feats_df)
trnRFEResults <-
myrun_rfe(glbObsTrn, indepVar, glbRFESizes[["Final"]])
if (!isTRUE(all.equal(sort(predictors(trnRFEResults)),
sort(predictors(glbRFEResults))))) {
print("Diffs predictors(trnRFEResults) vs. predictors(glbRFEResults):")
print(setdiff(predictors(trnRFEResults), predictors(glbRFEResults)))
print("Diffs predictors(glbRFEResults) vs. predictors(trnRFEResults):")
print(setdiff(predictors(glbRFEResults), predictors(trnRFEResults)))
}
}
if (grepl("Ensemble", glbMdlSelId)) {
# Find which models are relevant
mdlimp_df <- subset(myget_feats_importance(glb_sel_mdl), imp > 5)
# Fit selected models on glbObsTrn
for (mdl_id in gsub(".prob", "",
gsub(mygetPredictIds(glb_rsp_var)$value, "", row.names(mdlimp_df), fixed = TRUE),
fixed = TRUE)) {
mdl_id_components <- unlist(strsplit(mdl_id, "[.]"))
mdlIdPfx <- paste0(c(head(mdl_id_components, -1), "Train"),
collapse = ".")
if (grepl("RFE\\.X\\.", mdlIdPfx))
mdlIndepVars <- myadjustInteractionFeats(glb_feats_df, myextract_actual_feats(
predictors(trnRFEResults))) else
mdlIndepVars <- trim(unlist(
strsplit(glb_models_df[glb_models_df$id == mdl_id, "feats"], "[,]")))
ret_lst <-
myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = mdlIdPfx,
type = glb_model_type, tune.df = glbMdlTuneParams,
trainControl.method = "repeatedcv",
trainControl.number = glb_rcv_n_folds,
trainControl.repeats = glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method = tail(mdl_id_components, 1))),
indepVar = mdlIndepVars,
rsp_var = glb_rsp_var,
fit_df = glbObsTrn, OOB_df = NULL)
glbObsTrn <- glb_get_predictions(df = glbObsTrn,
mdl_id = tail(glb_models_df$id, 1),
rsp_var = glb_rsp_var,
prob_threshold_def =
subset(glb_models_df, id == mdl_id)$opt.prob.threshold.OOB)
glbObsNew <- glb_get_predictions(df = glbObsNew,
mdl_id = tail(glb_models_df$id, 1),
rsp_var = glb_rsp_var,
prob_threshold_def =
subset(glb_models_df, id == mdl_id)$opt.prob.threshold.OOB)
}
}
# "Final" model
if ((model_method <- glb_sel_mdl$method) == "custom")
# get actual method from the mdl_id
model_method <- tail(unlist(strsplit(glbMdlSelId, "[.]")), 1)
if (grepl("Ensemble", glbMdlSelId)) {
# Find which models are relevant
mdlimp_df <- subset(myget_feats_importance(glb_sel_mdl), imp > 5)
if (glb_is_classification && glb_is_binomial)
indepVar <- gsub("(.*)\\.(.*)\\.prob", "\\1\\.Train\\.\\2\\.prob",
row.names(mdlimp_df)) else
indepVar <- gsub("(.*)\\.(.*)", "\\1\\.Train\\.\\2",
row.names(mdlimp_df))
} else
if (grepl("RFE.X", glbMdlSelId, fixed = TRUE)) {
# indepVar <- myextract_actual_feats(predictors(trnRFEResults))
indepVar <- myextract_actual_feats(predictors(glbRFEResults))
} else indepVar <-
trim(unlist(strsplit(glb_models_df[glb_models_df$id ==
glbMdlSelId
, "feats"], "[,]")))
# if (!is.null(glbMdlPreprocMethods) &&
# ((match_pos <- regexpr(gsub(".", "\\.",
# paste(glbMdlPreprocMethods, collapse = "|"),
# fixed = TRUE), glbMdlSelId)) != -1))
# ths_preProcess <- str_sub(glbMdlSelId, match_pos,
# match_pos + attr(match_pos, "match.length") - 1) else
# ths_preProcess <- NULL
# mdl_id_pfx <- ifelse(grepl("Ensemble", glbMdlSelId),
# "Final.Ensemble", "Final")
mdl_id_pfx <- paste("Final", myparseMdlId(glbMdlSelId)$family, sep = ".")
trnobs_df <- glbObsTrn
if (!is.null(glbObsTrnOutliers[[mdl_id_pfx]])) {
trnobs_df <- glbObsTrn[!(glbObsTrn[, glbFeatsId] %in% glbObsTrnOutliers[[mdl_id_pfx]]), ]
print(sprintf("Outliers removed: %d", nrow(glbObsTrn) - nrow(trnobs_df)))
print(setdiff(glbObsTrn[, glbFeatsId], trnobs_df[, glbFeatsId]))
}
# Force fitting of Final.glm to identify outliers
method_vctr <- unique(c(myparseMdlId(glbMdlSelId)$alg, glbMdlFamilies[["Final"]]))
for (method in method_vctr) {
#source("caret_nominalTrainWorkflow.R")
# glmnet requires at least 2 indep vars
if ((length(indepVar) == 1) && (method %in% "glmnet"))
next
ret_lst <-
myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = mdl_id_pfx,
type = glb_model_type, trainControl.method = "repeatedcv",
trainControl.number = glb_rcv_n_folds,
trainControl.repeats = glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
trainControl.allowParallel = glbMdlAllowParallel,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method = method,
train.preProcess = myparseMdlId(glbMdlSelId)$preProcess)),
indepVar = indepVar, rsp_var = glb_rsp_var,
fit_df = trnobs_df, OOB_df = NULL)
if ((length(method_vctr) == 1) || (method != "glm")) {
glb_fin_mdl <- glb_models_lst[[length(glb_models_lst)]]
glbMdlFinId <- glb_models_df[length(glb_models_lst), "id"]
}
}
}## +(rfe) fit Fold1.Rep1 size: 247## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold1.Rep1 size: 247
## +(rfe) imp Fold1.Rep1
## -(rfe) imp Fold1.Rep1
## +(rfe) fit Fold1.Rep1 size: 128
## -(rfe) fit Fold1.Rep1 size: 128
## +(rfe) fit Fold1.Rep1 size: 96
## -(rfe) fit Fold1.Rep1 size: 96
## +(rfe) fit Fold1.Rep1 size: 64
## -(rfe) fit Fold1.Rep1 size: 64
## +(rfe) fit Fold1.Rep1 size: 56
## -(rfe) fit Fold1.Rep1 size: 56
## +(rfe) fit Fold1.Rep1 size: 52
## -(rfe) fit Fold1.Rep1 size: 52
## +(rfe) fit Fold1.Rep1 size: 51
## -(rfe) fit Fold1.Rep1 size: 51
## +(rfe) fit Fold1.Rep1 size: 50
## -(rfe) fit Fold1.Rep1 size: 50
## +(rfe) fit Fold1.Rep1 size: 49
## -(rfe) fit Fold1.Rep1 size: 49
## +(rfe) fit Fold1.Rep1 size: 48
## -(rfe) fit Fold1.Rep1 size: 48
## +(rfe) fit Fold1.Rep1 size: 46
## -(rfe) fit Fold1.Rep1 size: 46
## +(rfe) fit Fold1.Rep1 size: 44
## -(rfe) fit Fold1.Rep1 size: 44
## +(rfe) fit Fold1.Rep1 size: 40
## -(rfe) fit Fold1.Rep1 size: 40
## +(rfe) fit Fold1.Rep1 size: 32
## -(rfe) fit Fold1.Rep1 size: 32
## +(rfe) fit Fold1.Rep1 size: 16
## -(rfe) fit Fold1.Rep1 size: 16
## +(rfe) fit Fold1.Rep1 size: 8
## -(rfe) fit Fold1.Rep1 size: 8
## +(rfe) fit Fold2.Rep1 size: 247## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold2.Rep1 size: 247
## +(rfe) imp Fold2.Rep1
## -(rfe) imp Fold2.Rep1
## +(rfe) fit Fold2.Rep1 size: 128
## -(rfe) fit Fold2.Rep1 size: 128
## +(rfe) fit Fold2.Rep1 size: 96
## -(rfe) fit Fold2.Rep1 size: 96
## +(rfe) fit Fold2.Rep1 size: 64
## -(rfe) fit Fold2.Rep1 size: 64
## +(rfe) fit Fold2.Rep1 size: 56
## -(rfe) fit Fold2.Rep1 size: 56
## +(rfe) fit Fold2.Rep1 size: 52
## -(rfe) fit Fold2.Rep1 size: 52
## +(rfe) fit Fold2.Rep1 size: 51
## -(rfe) fit Fold2.Rep1 size: 51
## +(rfe) fit Fold2.Rep1 size: 50
## -(rfe) fit Fold2.Rep1 size: 50
## +(rfe) fit Fold2.Rep1 size: 49
## -(rfe) fit Fold2.Rep1 size: 49
## +(rfe) fit Fold2.Rep1 size: 48
## -(rfe) fit Fold2.Rep1 size: 48
## +(rfe) fit Fold2.Rep1 size: 46
## -(rfe) fit Fold2.Rep1 size: 46
## +(rfe) fit Fold2.Rep1 size: 44
## -(rfe) fit Fold2.Rep1 size: 44
## +(rfe) fit Fold2.Rep1 size: 40
## -(rfe) fit Fold2.Rep1 size: 40
## +(rfe) fit Fold2.Rep1 size: 32
## -(rfe) fit Fold2.Rep1 size: 32
## +(rfe) fit Fold2.Rep1 size: 16
## -(rfe) fit Fold2.Rep1 size: 16
## +(rfe) fit Fold2.Rep1 size: 8
## -(rfe) fit Fold2.Rep1 size: 8
## +(rfe) fit Fold3.Rep1 size: 247## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold3.Rep1 size: 247
## +(rfe) imp Fold3.Rep1
## -(rfe) imp Fold3.Rep1
## +(rfe) fit Fold3.Rep1 size: 128
## -(rfe) fit Fold3.Rep1 size: 128
## +(rfe) fit Fold3.Rep1 size: 96
## -(rfe) fit Fold3.Rep1 size: 96
## +(rfe) fit Fold3.Rep1 size: 64
## -(rfe) fit Fold3.Rep1 size: 64
## +(rfe) fit Fold3.Rep1 size: 56
## -(rfe) fit Fold3.Rep1 size: 56
## +(rfe) fit Fold3.Rep1 size: 52
## -(rfe) fit Fold3.Rep1 size: 52
## +(rfe) fit Fold3.Rep1 size: 51
## -(rfe) fit Fold3.Rep1 size: 51
## +(rfe) fit Fold3.Rep1 size: 50
## -(rfe) fit Fold3.Rep1 size: 50
## +(rfe) fit Fold3.Rep1 size: 49
## -(rfe) fit Fold3.Rep1 size: 49
## +(rfe) fit Fold3.Rep1 size: 48
## -(rfe) fit Fold3.Rep1 size: 48
## +(rfe) fit Fold3.Rep1 size: 46
## -(rfe) fit Fold3.Rep1 size: 46
## +(rfe) fit Fold3.Rep1 size: 44
## -(rfe) fit Fold3.Rep1 size: 44
## +(rfe) fit Fold3.Rep1 size: 40
## -(rfe) fit Fold3.Rep1 size: 40
## +(rfe) fit Fold3.Rep1 size: 32
## -(rfe) fit Fold3.Rep1 size: 32
## +(rfe) fit Fold3.Rep1 size: 16
## -(rfe) fit Fold3.Rep1 size: 16
## +(rfe) fit Fold3.Rep1 size: 8
## -(rfe) fit Fold3.Rep1 size: 8
## +(rfe) fit Fold1.Rep2 size: 247## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold1.Rep2 size: 247
## +(rfe) imp Fold1.Rep2
## -(rfe) imp Fold1.Rep2
## +(rfe) fit Fold1.Rep2 size: 128
## -(rfe) fit Fold1.Rep2 size: 128
## +(rfe) fit Fold1.Rep2 size: 96
## -(rfe) fit Fold1.Rep2 size: 96
## +(rfe) fit Fold1.Rep2 size: 64
## -(rfe) fit Fold1.Rep2 size: 64
## +(rfe) fit Fold1.Rep2 size: 56
## -(rfe) fit Fold1.Rep2 size: 56
## +(rfe) fit Fold1.Rep2 size: 52
## -(rfe) fit Fold1.Rep2 size: 52
## +(rfe) fit Fold1.Rep2 size: 51
## -(rfe) fit Fold1.Rep2 size: 51
## +(rfe) fit Fold1.Rep2 size: 50
## -(rfe) fit Fold1.Rep2 size: 50
## +(rfe) fit Fold1.Rep2 size: 49
## -(rfe) fit Fold1.Rep2 size: 49
## +(rfe) fit Fold1.Rep2 size: 48
## -(rfe) fit Fold1.Rep2 size: 48
## +(rfe) fit Fold1.Rep2 size: 46
## -(rfe) fit Fold1.Rep2 size: 46
## +(rfe) fit Fold1.Rep2 size: 44
## -(rfe) fit Fold1.Rep2 size: 44
## +(rfe) fit Fold1.Rep2 size: 40
## -(rfe) fit Fold1.Rep2 size: 40
## +(rfe) fit Fold1.Rep2 size: 32
## -(rfe) fit Fold1.Rep2 size: 32
## +(rfe) fit Fold1.Rep2 size: 16
## -(rfe) fit Fold1.Rep2 size: 16
## +(rfe) fit Fold1.Rep2 size: 8
## -(rfe) fit Fold1.Rep2 size: 8
## +(rfe) fit Fold2.Rep2 size: 247## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold2.Rep2 size: 247
## +(rfe) imp Fold2.Rep2
## -(rfe) imp Fold2.Rep2
## +(rfe) fit Fold2.Rep2 size: 128
## -(rfe) fit Fold2.Rep2 size: 128
## +(rfe) fit Fold2.Rep2 size: 96
## -(rfe) fit Fold2.Rep2 size: 96
## +(rfe) fit Fold2.Rep2 size: 64
## -(rfe) fit Fold2.Rep2 size: 64
## +(rfe) fit Fold2.Rep2 size: 56
## -(rfe) fit Fold2.Rep2 size: 56
## +(rfe) fit Fold2.Rep2 size: 52
## -(rfe) fit Fold2.Rep2 size: 52
## +(rfe) fit Fold2.Rep2 size: 51
## -(rfe) fit Fold2.Rep2 size: 51
## +(rfe) fit Fold2.Rep2 size: 50
## -(rfe) fit Fold2.Rep2 size: 50
## +(rfe) fit Fold2.Rep2 size: 49
## -(rfe) fit Fold2.Rep2 size: 49
## +(rfe) fit Fold2.Rep2 size: 48
## -(rfe) fit Fold2.Rep2 size: 48
## +(rfe) fit Fold2.Rep2 size: 46
## -(rfe) fit Fold2.Rep2 size: 46
## +(rfe) fit Fold2.Rep2 size: 44
## -(rfe) fit Fold2.Rep2 size: 44
## +(rfe) fit Fold2.Rep2 size: 40
## -(rfe) fit Fold2.Rep2 size: 40
## +(rfe) fit Fold2.Rep2 size: 32
## -(rfe) fit Fold2.Rep2 size: 32
## +(rfe) fit Fold2.Rep2 size: 16
## -(rfe) fit Fold2.Rep2 size: 16
## +(rfe) fit Fold2.Rep2 size: 8
## -(rfe) fit Fold2.Rep2 size: 8
## +(rfe) fit Fold3.Rep2 size: 247## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold3.Rep2 size: 247
## +(rfe) imp Fold3.Rep2
## -(rfe) imp Fold3.Rep2
## +(rfe) fit Fold3.Rep2 size: 128
## -(rfe) fit Fold3.Rep2 size: 128
## +(rfe) fit Fold3.Rep2 size: 96
## -(rfe) fit Fold3.Rep2 size: 96
## +(rfe) fit Fold3.Rep2 size: 64
## -(rfe) fit Fold3.Rep2 size: 64
## +(rfe) fit Fold3.Rep2 size: 56
## -(rfe) fit Fold3.Rep2 size: 56
## +(rfe) fit Fold3.Rep2 size: 52
## -(rfe) fit Fold3.Rep2 size: 52
## +(rfe) fit Fold3.Rep2 size: 51
## -(rfe) fit Fold3.Rep2 size: 51
## +(rfe) fit Fold3.Rep2 size: 50
## -(rfe) fit Fold3.Rep2 size: 50
## +(rfe) fit Fold3.Rep2 size: 49
## -(rfe) fit Fold3.Rep2 size: 49
## +(rfe) fit Fold3.Rep2 size: 48
## -(rfe) fit Fold3.Rep2 size: 48
## +(rfe) fit Fold3.Rep2 size: 46
## -(rfe) fit Fold3.Rep2 size: 46
## +(rfe) fit Fold3.Rep2 size: 44
## -(rfe) fit Fold3.Rep2 size: 44
## +(rfe) fit Fold3.Rep2 size: 40
## -(rfe) fit Fold3.Rep2 size: 40
## +(rfe) fit Fold3.Rep2 size: 32
## -(rfe) fit Fold3.Rep2 size: 32
## +(rfe) fit Fold3.Rep2 size: 16
## -(rfe) fit Fold3.Rep2 size: 16
## +(rfe) fit Fold3.Rep2 size: 8
## -(rfe) fit Fold3.Rep2 size: 8
## +(rfe) fit Fold1.Rep3 size: 247## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold1.Rep3 size: 247
## +(rfe) imp Fold1.Rep3
## -(rfe) imp Fold1.Rep3
## +(rfe) fit Fold1.Rep3 size: 128
## -(rfe) fit Fold1.Rep3 size: 128
## +(rfe) fit Fold1.Rep3 size: 96
## -(rfe) fit Fold1.Rep3 size: 96
## +(rfe) fit Fold1.Rep3 size: 64
## -(rfe) fit Fold1.Rep3 size: 64
## +(rfe) fit Fold1.Rep3 size: 56
## -(rfe) fit Fold1.Rep3 size: 56
## +(rfe) fit Fold1.Rep3 size: 52
## -(rfe) fit Fold1.Rep3 size: 52
## +(rfe) fit Fold1.Rep3 size: 51
## -(rfe) fit Fold1.Rep3 size: 51
## +(rfe) fit Fold1.Rep3 size: 50
## -(rfe) fit Fold1.Rep3 size: 50
## +(rfe) fit Fold1.Rep3 size: 49
## -(rfe) fit Fold1.Rep3 size: 49
## +(rfe) fit Fold1.Rep3 size: 48
## -(rfe) fit Fold1.Rep3 size: 48
## +(rfe) fit Fold1.Rep3 size: 46
## -(rfe) fit Fold1.Rep3 size: 46
## +(rfe) fit Fold1.Rep3 size: 44
## -(rfe) fit Fold1.Rep3 size: 44
## +(rfe) fit Fold1.Rep3 size: 40
## -(rfe) fit Fold1.Rep3 size: 40
## +(rfe) fit Fold1.Rep3 size: 32
## -(rfe) fit Fold1.Rep3 size: 32
## +(rfe) fit Fold1.Rep3 size: 16
## -(rfe) fit Fold1.Rep3 size: 16
## +(rfe) fit Fold1.Rep3 size: 8
## -(rfe) fit Fold1.Rep3 size: 8
## +(rfe) fit Fold2.Rep3 size: 247## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold2.Rep3 size: 247
## +(rfe) imp Fold2.Rep3
## -(rfe) imp Fold2.Rep3
## +(rfe) fit Fold2.Rep3 size: 128
## -(rfe) fit Fold2.Rep3 size: 128
## +(rfe) fit Fold2.Rep3 size: 96
## -(rfe) fit Fold2.Rep3 size: 96
## +(rfe) fit Fold2.Rep3 size: 64
## -(rfe) fit Fold2.Rep3 size: 64
## +(rfe) fit Fold2.Rep3 size: 56
## -(rfe) fit Fold2.Rep3 size: 56
## +(rfe) fit Fold2.Rep3 size: 52
## -(rfe) fit Fold2.Rep3 size: 52
## +(rfe) fit Fold2.Rep3 size: 51
## -(rfe) fit Fold2.Rep3 size: 51
## +(rfe) fit Fold2.Rep3 size: 50
## -(rfe) fit Fold2.Rep3 size: 50
## +(rfe) fit Fold2.Rep3 size: 49
## -(rfe) fit Fold2.Rep3 size: 49
## +(rfe) fit Fold2.Rep3 size: 48
## -(rfe) fit Fold2.Rep3 size: 48
## +(rfe) fit Fold2.Rep3 size: 46
## -(rfe) fit Fold2.Rep3 size: 46
## +(rfe) fit Fold2.Rep3 size: 44
## -(rfe) fit Fold2.Rep3 size: 44
## +(rfe) fit Fold2.Rep3 size: 40
## -(rfe) fit Fold2.Rep3 size: 40
## +(rfe) fit Fold2.Rep3 size: 32
## -(rfe) fit Fold2.Rep3 size: 32
## +(rfe) fit Fold2.Rep3 size: 16
## -(rfe) fit Fold2.Rep3 size: 16
## +(rfe) fit Fold2.Rep3 size: 8
## -(rfe) fit Fold2.Rep3 size: 8
## +(rfe) fit Fold3.Rep3 size: 247## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold3.Rep3 size: 247
## +(rfe) imp Fold3.Rep3
## -(rfe) imp Fold3.Rep3
## +(rfe) fit Fold3.Rep3 size: 128
## -(rfe) fit Fold3.Rep3 size: 128
## +(rfe) fit Fold3.Rep3 size: 96
## -(rfe) fit Fold3.Rep3 size: 96
## +(rfe) fit Fold3.Rep3 size: 64
## -(rfe) fit Fold3.Rep3 size: 64
## +(rfe) fit Fold3.Rep3 size: 56
## -(rfe) fit Fold3.Rep3 size: 56
## +(rfe) fit Fold3.Rep3 size: 52
## -(rfe) fit Fold3.Rep3 size: 52
## +(rfe) fit Fold3.Rep3 size: 51
## -(rfe) fit Fold3.Rep3 size: 51
## +(rfe) fit Fold3.Rep3 size: 50
## -(rfe) fit Fold3.Rep3 size: 50
## +(rfe) fit Fold3.Rep3 size: 49
## -(rfe) fit Fold3.Rep3 size: 49
## +(rfe) fit Fold3.Rep3 size: 48
## -(rfe) fit Fold3.Rep3 size: 48
## +(rfe) fit Fold3.Rep3 size: 46
## -(rfe) fit Fold3.Rep3 size: 46
## +(rfe) fit Fold3.Rep3 size: 44
## -(rfe) fit Fold3.Rep3 size: 44
## +(rfe) fit Fold3.Rep3 size: 40
## -(rfe) fit Fold3.Rep3 size: 40
## +(rfe) fit Fold3.Rep3 size: 32
## -(rfe) fit Fold3.Rep3 size: 32
## +(rfe) fit Fold3.Rep3 size: 16
## -(rfe) fit Fold3.Rep3 size: 16
## +(rfe) fit Fold3.Rep3 size: 8
## -(rfe) fit Fold3.Rep3 size: 8
##
## Recursive feature selection
##
## Outer resampling method: Cross-Validated (3 fold, repeated 3 times)
##
## Resampling performance over subset size:
##
## Variables Accuracy Kappa AccuracySD KappaSD Selected
## 8 0.6061 0.2027 0.015046 0.02955
## 16 0.6066 0.2021 0.016324 0.03294
## 32 0.6103 0.2102 0.014275 0.02881
## 40 0.6142 0.2185 0.019295 0.03900
## 44 0.6151 0.2206 0.017705 0.03573
## 46 0.6147 0.2198 0.015989 0.03232
## 48 0.6153 0.2210 0.018660 0.03771
## 49 0.6164 0.2232 0.019028 0.03847 *
## 50 0.6154 0.2214 0.018224 0.03687
## 51 0.6152 0.2209 0.017556 0.03546
## 52 0.6139 0.2183 0.017950 0.03626
## 56 0.6130 0.2167 0.014393 0.02893
## 64 0.6149 0.2206 0.016610 0.03353
## 96 0.6124 0.2169 0.012169 0.02474
## 128 0.6100 0.2130 0.009981 0.02037
## 247 0.6074 0.2110 0.008634 0.01737
##
## The top 5 variables (out of 49):
## Q109244.fctrNo, Q115611.fctrYes, Q113181.fctrYes, Gender.fctrM, Q98197.fctrYes
##
## [1] "Q109244.fctrNo" "Q115611.fctrYes"
## [3] "Q113181.fctrYes" "Gender.fctrM"
## [5] "Q98197.fctrYes" "Q101163.fctrDad"
## [7] "Hhold.fctrMKy" "Q120472.fctrScience"
## [9] "Q120379.fctrNo" "Q105840.fctrNo"
## [11] "Q98869.fctrYes" "Q99480.fctrYes"
## [13] "Q106272.fctrYes" "Q109244.fctrNo:.clusterid.fctr3"
## [15] "Q119851.fctrNo" "Q116881.fctrRight"
## [17] "Q110740.fctrPC" "Q115899.fctrMe"
## [19] "Q106042.fctrNo" "Income.fctr.Q"
## [21] "Q120014.fctrYes" "Q108855.fctrYes!"
## [23] "Q101596.fctrYes" "Q121699.fctrNo"
## [25] "Q121699.fctrYes" "Q102089.fctrOwn"
## [27] "Q107869.fctrYes" "Q118232.fctrPr"
## [29] "Q112478.fctrNo" "Edn.fctr^7"
## [31] "Q120650.fctrYes" "Q100680.fctrNo"
## [33] "Q120012.fctrNo" "Q106389.fctrNo"
## [35] "Q115390.fctrNo" "Q118237.fctrNo"
## [37] "Q118892.fctrNo" "Q121011.fctrNo"
## [39] "Q111220.fctrNo" "Q115195.fctrNo"
## [41] "Q108342.fctrIn-person" "Q116448.fctrNo"
## [43] "Q108617.fctrNo" "Q116953.fctrNo"
## [45] "Q124122.fctrNo" "Q119650.fctrGiving"
## [47] "Q122771.fctrPt" "Q106388.fctrYes"
## [49] "Q120194.fctrTry first"
## [1] "Diffs predictors(trnRFEResults) vs. predictors(glbRFEResults):"
## character(0)
## [1] "Diffs predictors(glbRFEResults) vs. predictors(trnRFEResults):"
## [1] "Edn.fctr^6" "Q96024.fctrYes"
## [3] "Q123621.fctrYes" "Q113583.fctrTalk"
## [5] "Q122770.fctrYes" "Q121700.fctrNo"
## [7] "Q114152.fctrYes" "Q106997.fctrGr"
## [9] "Q98078.fctrNo" "Q112270.fctrYes"
## [11] "Q111848.fctrNo" "Q100689.fctrNo"
## [13] "Income.fctr^4" "Q118233.fctrNo"
## [15] "Q99982.fctrNope" "Q106993.fctrYes"
## [17] "Q102906.fctrNo" "Q123464.fctrNo"
## [19] "Q116797.fctrNo" "Q104996.fctrNo"
## [21] "Q109367.fctrNo" "Q99716.fctrNo"
## [23] "Q117186.fctrHot headed" "Q102674.fctrNo"
## [25] "Q118117.fctrYes" "Q122769.fctrYes"
## [27] "Edn.fctr.C" "Q108754.fctrNo"
## [29] "Q116441.fctrYes" "YOB.Age.fctr^7"
## [31] "Q105655.fctrYes" "Q108856.fctrSocialize"
## [33] "YOB.Age.fctr^8" "Q120978.fctrNo"
## [35] "Q101162.fctrPessimist" "Q122769.fctrNo"
## [37] "Q119334.fctrNo" "Q102289.fctrYes"
## [39] "Q111580.fctrSupportive" "Q112512.fctrYes"
## [41] "Q99581.fctrNo" "Q122120.fctrNo"
## [43] "Income.fctr.L" "Q116197.fctrP.M."
## [45] "Q114386.fctrMysterious" "Q117193.fctrStandard hours"
## [47] "Q114961.fctrNo" "Q115777.fctrStart"
## [49] "YOB.Age.fctr(30,35]:YOB.Age.dff" "Q100010.fctrYes"
## [51] "Q122120.fctrYes" "Q113584.fctrTechnology"
## [53] "Q119334.fctrYes" "Q98578.fctrYes"
## [55] "Q102687.fctrNo" "Q100562.fctrNo"
## [57] "Q116601.fctrNo" "Q116197.fctrA.M."
## [59] "Q103293.fctrYes" "Q108950.fctrCautious"
## [61] "Q98578.fctrNo" "Q103293.fctrNo"
## [63] "Q115602.fctrYes" "YOB.Age.fctr(40,50]:YOB.Age.dff"
## [65] "Q107491.fctrNo" "Q114748.fctrYes"
## [67] "Q115777.fctrEnd" "Q114517.fctrYes"
## [69] "Edn.fctr^5" "Q108856.fctrSpace"
## [71] "YOB.Age.fctr^5" "Q115610.fctrYes"
## [73] "Income.fctr.C" "Q109244.fctrNo:.clusterid.fctr2"
## [75] "Q124742.fctrYes" "Q108343.fctrNo"
## [77] "Income.fctr^5" "Q105655.fctrNo"
## [79] "Q111580.fctrDemanding" "Q115610.fctrNo"
## [81] "Q108343.fctrYes" "Q113992.fctrNo"
## [1] "\nmyfit_mdl: enter: 0.001000 secs"
## [1] "fitting model: Final.RFE.X#zv.pca#rcv#glmnet"
## [1] " indepVar: Q109244.fctr,Q115611.fctr,Q113181.fctr,Gender.fctr,Q98197.fctr,Q101163.fctr,Q120379.fctr,Hhold.fctr,Q105840.fctr,.clusterid.fctr,Q120472.fctr,Q106272.fctr,Q115899.fctr,Q119851.fctr,Q99480.fctr,Q106042.fctr,Q98869.fctr,Q101596.fctr,Q110740.fctr,Q102089.fctr,Q120014.fctr,Q120650.fctr,Q108855.fctr,Edn.fctr,Q118892.fctr,Q116881.fctr,Q100680.fctr,Q108342.fctr,Q107869.fctr,Q112478.fctr,Q115195.fctr,Q121699.fctr,Q106388.fctr,Q106389.fctr,Q120012.fctr,Q116448.fctr,Q119650.fctr,Q96024.fctr,Q118237.fctr,Income.fctr,Q123621.fctr,Q108617.fctr,Q113583.fctr,Q118232.fctr,Q120194.fctr,Q122770.fctr,Q121700.fctr,Q124122.fctr,Q114152.fctr,Q122771.fctr,Q106997.fctr,Q98078.fctr,Q112270.fctr,Q116953.fctr,Q111848.fctr,Q100689.fctr,Q118233.fctr,Q99982.fctr,Q106993.fctr,Q115390.fctr,Q102906.fctr,Q123464.fctr,Q116797.fctr,Q104996.fctr,Q109367.fctr,Q111220.fctr,Q99716.fctr,Q117186.fctr,Q102674.fctr,Q118117.fctr,Q122769.fctr,Q108754.fctr,Q121011.fctr,Q116441.fctr,YOB.Age.fctr,Q105655.fctr,Q108856.fctr,Q120978.fctr,Q101162.fctr,Q119334.fctr,Q102289.fctr,Q111580.fctr,Q112512.fctr,Q99581.fctr,Q122120.fctr,Q116197.fctr,Q114386.fctr,Q117193.fctr,Q114961.fctr,Q115777.fctr,YOB.Age.dff,Q100010.fctr,Q113584.fctr,Q98578.fctr,Q102687.fctr,Q100562.fctr,Q116601.fctr,Q103293.fctr,Q108950.fctr,Q115602.fctr,Q107491.fctr,Q114748.fctr,Q114517.fctr,Q115610.fctr,Q124742.fctr,Q108343.fctr,Q113992.fctr"
## [1] "myfit_mdl: setup complete: 0.726000 secs"
## + Fold1.Rep1: alpha=0.100, lambda=0.04062
## - Fold1.Rep1: alpha=0.100, lambda=0.04062
## + Fold1.Rep1: alpha=0.325, lambda=0.04062
## - Fold1.Rep1: alpha=0.325, lambda=0.04062
## + Fold1.Rep1: alpha=0.550, lambda=0.04062
## - Fold1.Rep1: alpha=0.550, lambda=0.04062
## + Fold1.Rep1: alpha=0.775, lambda=0.04062
## - Fold1.Rep1: alpha=0.775, lambda=0.04062
## + Fold1.Rep1: alpha=1.000, lambda=0.04062
## - Fold1.Rep1: alpha=1.000, lambda=0.04062
## + Fold2.Rep1: alpha=0.100, lambda=0.04062
## - Fold2.Rep1: alpha=0.100, lambda=0.04062
## + Fold2.Rep1: alpha=0.325, lambda=0.04062
## - Fold2.Rep1: alpha=0.325, lambda=0.04062
## + Fold2.Rep1: alpha=0.550, lambda=0.04062
## - Fold2.Rep1: alpha=0.550, lambda=0.04062
## + Fold2.Rep1: alpha=0.775, lambda=0.04062
## - Fold2.Rep1: alpha=0.775, lambda=0.04062
## + Fold2.Rep1: alpha=1.000, lambda=0.04062
## - Fold2.Rep1: alpha=1.000, lambda=0.04062
## + Fold3.Rep1: alpha=0.100, lambda=0.04062
## - Fold3.Rep1: alpha=0.100, lambda=0.04062
## + Fold3.Rep1: alpha=0.325, lambda=0.04062
## - Fold3.Rep1: alpha=0.325, lambda=0.04062
## + Fold3.Rep1: alpha=0.550, lambda=0.04062
## - Fold3.Rep1: alpha=0.550, lambda=0.04062
## + Fold3.Rep1: alpha=0.775, lambda=0.04062
## - Fold3.Rep1: alpha=0.775, lambda=0.04062
## + Fold3.Rep1: alpha=1.000, lambda=0.04062
## - Fold3.Rep1: alpha=1.000, lambda=0.04062
## + Fold1.Rep2: alpha=0.100, lambda=0.04062
## - Fold1.Rep2: alpha=0.100, lambda=0.04062
## + Fold1.Rep2: alpha=0.325, lambda=0.04062
## - Fold1.Rep2: alpha=0.325, lambda=0.04062
## + Fold1.Rep2: alpha=0.550, lambda=0.04062
## - Fold1.Rep2: alpha=0.550, lambda=0.04062
## + Fold1.Rep2: alpha=0.775, lambda=0.04062
## - Fold1.Rep2: alpha=0.775, lambda=0.04062
## + Fold1.Rep2: alpha=1.000, lambda=0.04062
## - Fold1.Rep2: alpha=1.000, lambda=0.04062
## + Fold2.Rep2: alpha=0.100, lambda=0.04062
## - Fold2.Rep2: alpha=0.100, lambda=0.04062
## + Fold2.Rep2: alpha=0.325, lambda=0.04062
## - Fold2.Rep2: alpha=0.325, lambda=0.04062
## + Fold2.Rep2: alpha=0.550, lambda=0.04062
## - Fold2.Rep2: alpha=0.550, lambda=0.04062
## + Fold2.Rep2: alpha=0.775, lambda=0.04062
## - Fold2.Rep2: alpha=0.775, lambda=0.04062
## + Fold2.Rep2: alpha=1.000, lambda=0.04062
## - Fold2.Rep2: alpha=1.000, lambda=0.04062
## + Fold3.Rep2: alpha=0.100, lambda=0.04062
## - Fold3.Rep2: alpha=0.100, lambda=0.04062
## + Fold3.Rep2: alpha=0.325, lambda=0.04062
## - Fold3.Rep2: alpha=0.325, lambda=0.04062
## + Fold3.Rep2: alpha=0.550, lambda=0.04062
## - Fold3.Rep2: alpha=0.550, lambda=0.04062
## + Fold3.Rep2: alpha=0.775, lambda=0.04062
## - Fold3.Rep2: alpha=0.775, lambda=0.04062
## + Fold3.Rep2: alpha=1.000, lambda=0.04062
## - Fold3.Rep2: alpha=1.000, lambda=0.04062
## + Fold1.Rep3: alpha=0.100, lambda=0.04062
## - Fold1.Rep3: alpha=0.100, lambda=0.04062
## + Fold1.Rep3: alpha=0.325, lambda=0.04062
## - Fold1.Rep3: alpha=0.325, lambda=0.04062
## + Fold1.Rep3: alpha=0.550, lambda=0.04062
## - Fold1.Rep3: alpha=0.550, lambda=0.04062
## + Fold1.Rep3: alpha=0.775, lambda=0.04062
## - Fold1.Rep3: alpha=0.775, lambda=0.04062
## + Fold1.Rep3: alpha=1.000, lambda=0.04062
## - Fold1.Rep3: alpha=1.000, lambda=0.04062
## + Fold2.Rep3: alpha=0.100, lambda=0.04062
## - Fold2.Rep3: alpha=0.100, lambda=0.04062
## + Fold2.Rep3: alpha=0.325, lambda=0.04062
## - Fold2.Rep3: alpha=0.325, lambda=0.04062
## + Fold2.Rep3: alpha=0.550, lambda=0.04062
## - Fold2.Rep3: alpha=0.550, lambda=0.04062
## + Fold2.Rep3: alpha=0.775, lambda=0.04062
## - Fold2.Rep3: alpha=0.775, lambda=0.04062
## + Fold2.Rep3: alpha=1.000, lambda=0.04062
## - Fold2.Rep3: alpha=1.000, lambda=0.04062
## + Fold3.Rep3: alpha=0.100, lambda=0.04062
## - Fold3.Rep3: alpha=0.100, lambda=0.04062
## + Fold3.Rep3: alpha=0.325, lambda=0.04062
## - Fold3.Rep3: alpha=0.325, lambda=0.04062
## + Fold3.Rep3: alpha=0.550, lambda=0.04062
## - Fold3.Rep3: alpha=0.550, lambda=0.04062
## + Fold3.Rep3: alpha=0.775, lambda=0.04062
## - Fold3.Rep3: alpha=0.775, lambda=0.04062
## + Fold3.Rep3: alpha=1.000, lambda=0.04062
## - Fold3.Rep3: alpha=1.000, lambda=0.04062
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 1, lambda = 0.00875 on full training set
## [1] "myfit_mdl: train complete: 94.424000 secs"## Warning in myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst
## = list(id.prefix = mdl_id_pfx, : model's bestTune found at an extreme of
## tuneGrid for parameter: alpha
## Length Class Mode
## a0 64 -none- numeric
## beta 9152 dgCMatrix S4
## df 64 -none- numeric
## dim 2 -none- numeric
## lambda 64 -none- numeric
## dev.ratio 64 -none- numeric
## nulldev 1 -none- numeric
## npasses 1 -none- numeric
## jerr 1 -none- numeric
## offset 1 -none- logical
## classnames 2 -none- character
## call 5 -none- call
## nobs 1 -none- numeric
## lambdaOpt 1 -none- numeric
## xNames 143 -none- character
## problemType 1 -none- character
## tuneValue 2 data.frame list
## obsLevels 2 -none- character
## [1] "min lambda > lambdaOpt:"
## (Intercept) PC2 PC3 PC4 PC5
## -0.1327006681 0.0210662656 -0.0877888193 0.0417899970 -0.1042532934
## PC6 PC7 PC9 PC10 PC12
## -0.0302547417 0.2100782515 0.0228261127 0.1055386942 0.0909002772
## PC14 PC15 PC16 PC20 PC23
## -0.0359218384 -0.0866939426 -0.0284090500 0.0021295865 0.0400977194
## PC27 PC28 PC31 PC37 PC38
## -0.0237693776 0.0202106385 -0.0500965637 0.0031020807 -0.0401755947
## PC42 PC56 PC61 PC62 PC66
## -0.0348872995 -0.0408258157 0.0126715178 -0.0130830641 -0.0214020261
## PC71 PC72 PC74 PC77 PC86
## -0.0053840964 0.0059595427 0.0025397394 -0.0458931313 -0.0343981132
## PC88 PC91 PC92 PC94 PC97
## 0.0406823986 0.0077176109 -0.0904246271 0.0176968566 -0.0609112378
## PC100 PC105 PC106 PC107 PC108
## -0.0053485006 -0.0006374331 0.0091673919 0.0172500586 0.0122925023
## PC110 PC113 PC120 PC122 PC128
## -0.0026630849 -0.0435850118 -0.0241128550 0.0185458332 0.0105287421
## PC133 PC139
## 0.0141463690 -0.0176353383
## [1] "max lambda < lambdaOpt:"
## (Intercept) PC2 PC3 PC4 PC5
## -0.1338208543 0.0224823418 -0.0897897766 0.0437228080 -0.1065667129
## PC6 PC7 PC9 PC10 PC12
## -0.0322826064 0.2131293784 0.0251385795 0.1083202536 0.0939545933
## PC14 PC15 PC16 PC19 PC20
## -0.0386115724 -0.0898848758 -0.0312137507 -0.0020613180 0.0050243113
## PC23 PC27 PC28 PC31 PC37
## 0.0430832242 -0.0267243679 0.0231676453 -0.0532371992 0.0061978493
## PC38 PC42 PC43 PC44 PC54
## -0.0433863135 -0.0381792566 0.0022232839 0.0001705535 -0.0008040441
## PC55 PC56 PC61 PC62 PC66
## -0.0003430840 -0.0443717115 0.0160624914 -0.0165569057 -0.0248944454
## PC71 PC72 PC74 PC77 PC83
## -0.0087704273 0.0095943642 0.0059444800 -0.0494294437 -0.0028986573
## PC86 PC88 PC91 PC92 PC93
## -0.0380823412 0.0445153738 0.0115378223 -0.0944285261 -0.0017854522
## PC94 PC97 PC100 PC105 PC106
## 0.0215306095 -0.0650011868 -0.0089615381 -0.0045342285 0.0130677546
## PC107 PC108 PC110 PC113 PC120
## 0.0211373785 0.0162565151 -0.0066331818 -0.0478273643 -0.0283560147
## PC122 PC128 PC133 PC139 PC142
## 0.0229407357 0.0148733907 0.0189373916 -0.0231024419 0.0055967838
## [1] "myfit_mdl: train diagnostics complete: 100.545000 secs"## Loading required package: ROCR## Loading required package: gplots##
## Attaching package: 'gplots'## The following object is masked from 'package:stats':
##
## lowess
## Loading required package: sqldf## Loading required package: gsubfn## Loading required package: proto## Loading required package: RSQLite## Loading required package: DBI## Loading required package: tcltk
## Prediction
## Reference D R
## D 2121 830
## R 1108 1509
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 6.519397e-01 2.971401e-01 6.392607e-01 6.644592e-01 5.299928e-01
## AccuracyPValue McnemarPValue
## 6.144311e-76 3.129872e-10
## [1] "myfit_mdl: predict complete: 134.421000 secs"
## id
## 1 Final.RFE.X#zv.pca#rcv#glmnet
## feats
## 1 Q109244.fctr,Q115611.fctr,Q113181.fctr,Gender.fctr,Q98197.fctr,Q101163.fctr,Q120379.fctr,Hhold.fctr,Q105840.fctr,.clusterid.fctr,Q120472.fctr,Q106272.fctr,Q115899.fctr,Q119851.fctr,Q99480.fctr,Q106042.fctr,Q98869.fctr,Q101596.fctr,Q110740.fctr,Q102089.fctr,Q120014.fctr,Q120650.fctr,Q108855.fctr,Edn.fctr,Q118892.fctr,Q116881.fctr,Q100680.fctr,Q108342.fctr,Q107869.fctr,Q112478.fctr,Q115195.fctr,Q121699.fctr,Q106388.fctr,Q106389.fctr,Q120012.fctr,Q116448.fctr,Q119650.fctr,Q96024.fctr,Q118237.fctr,Income.fctr,Q123621.fctr,Q108617.fctr,Q113583.fctr,Q118232.fctr,Q120194.fctr,Q122770.fctr,Q121700.fctr,Q124122.fctr,Q114152.fctr,Q122771.fctr,Q106997.fctr,Q98078.fctr,Q112270.fctr,Q116953.fctr,Q111848.fctr,Q100689.fctr,Q118233.fctr,Q99982.fctr,Q106993.fctr,Q115390.fctr,Q102906.fctr,Q123464.fctr,Q116797.fctr,Q104996.fctr,Q109367.fctr,Q111220.fctr,Q99716.fctr,Q117186.fctr,Q102674.fctr,Q118117.fctr,Q122769.fctr,Q108754.fctr,Q121011.fctr,Q116441.fctr,YOB.Age.fctr,Q105655.fctr,Q108856.fctr,Q120978.fctr,Q101162.fctr,Q119334.fctr,Q102289.fctr,Q111580.fctr,Q112512.fctr,Q99581.fctr,Q122120.fctr,Q116197.fctr,Q114386.fctr,Q117193.fctr,Q114961.fctr,Q115777.fctr,YOB.Age.dff,Q100010.fctr,Q113584.fctr,Q98578.fctr,Q102687.fctr,Q100562.fctr,Q116601.fctr,Q103293.fctr,Q108950.fctr,Q115602.fctr,Q107491.fctr,Q114748.fctr,Q114517.fctr,Q115610.fctr,Q124742.fctr,Q108343.fctr,Q113992.fctr
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 25 93.59 2.123
## max.AUCpROC.fit max.Sens.fit max.Spec.fit max.AUCROCR.fit
## 1 0.6476769 0.7187394 0.5766144 0.7037082
## opt.prob.threshold.fit max.f.score.fit max.Accuracy.fit
## 1 0.5 0.6089588 0.6272737
## max.AccuracyLower.fit max.AccuracyUpper.fit max.Kappa.fit
## 1 0.6392607 0.6644592 0.2475136
## max.AccuracySD.fit max.KappaSD.fit
## 1 0.008937507 0.01753324
## [1] "myfit_mdl: exit: 134.438000 secs"rm(ret_lst)
glb_chunks_df <- myadd_chunk(glb_chunks_df, "fit.data.training", major.inc=FALSE)## label step_major step_minor label_minor bgn end
## 4 fit.data.training 2 0 0 339.957 517.207
## 5 fit.data.training 2 1 1 517.208 NA
## elapsed
## 4 177.251
## 5 NA#stop(here"); glb2Sav()
if (glb_is_classification && glb_is_binomial)
prob_threshold <- glb_models_df[glb_models_df$id == glbMdlSelId,
"opt.prob.threshold.OOB"] else
prob_threshold <- NULL
if (grepl("Ensemble", glbMdlFinId)) {
# Get predictions for each model in ensemble; Outliers that have been moved to OOB might not have been predicted yet
mdlEnsembleComps <- unlist(str_split(subset(glb_models_df,
id == glbMdlFinId)$feats, ","))
if (glb_is_classification && glb_is_binomial)
mdlEnsembleComps <- gsub("\\.prob$", "", mdlEnsembleComps)
mdlEnsembleComps <- gsub(paste0("^",
gsub(".", "\\.", mygetPredictIds(glb_rsp_var)$value, fixed = TRUE)),
"", mdlEnsembleComps)
for (mdl_id in mdlEnsembleComps) {
glbObsTrn <- glb_get_predictions(df = glbObsTrn, mdl_id = mdl_id,
rsp_var = glb_rsp_var,
prob_threshold_def = prob_threshold)
glbObsNew <- glb_get_predictions(df = glbObsNew, mdl_id = mdl_id,
rsp_var = glb_rsp_var,
prob_threshold_def = prob_threshold)
}
}
glbObsTrn <- glb_get_predictions(df = glbObsTrn, mdl_id = glbMdlFinId,
rsp_var = glb_rsp_var,
prob_threshold_def = prob_threshold)## Warning in glb_get_predictions(df = glbObsTrn, mdl_id = glbMdlFinId,
## rsp_var = glb_rsp_var, : Using default probability threshold: 0.5glb_featsimp_df <- myget_feats_importance(mdl=glb_fin_mdl,
featsimp_df=glb_featsimp_df)
#glb_featsimp_df[, paste0(glbMdlFinId, ".imp")] <- glb_featsimp_df$imp
print(glb_featsimp_df)## RFE.X.zv.pca.rcv.glmnet.imp Final.RFE.X.zv.pca.rcv.glmnet.imp
## PC7 0.000000e+00 100.00000000
## PC10 6.660214e+01 50.56857563
## PC5 7.995721e+01 49.83764917
## PC92 6.234515e+00 43.75600909
## PC12 1.279928e-01 43.72906314
## PC3 4.474537e+01 41.98091808
## PC15 6.321726e+01 41.77916319
## PC97 0.000000e+00 29.84358506
## PC31 0.000000e+00 24.48580379
## PC77 8.954286e+00 22.60589561
## PC113 2.752156e+00 21.70310070
## PC4 4.495763e+00 20.24378933
## PC88 2.477208e+01 20.22414706
## PC56 0.000000e+00 20.21582023
## PC38 0.000000e+00 19.82002416
## PC23 1.488607e+01 19.72359791
## PC14 4.398423e+01 17.67354338
## PC42 0.000000e+00 17.34459014
## PC86 0.000000e+00 17.21752030
## PC6 8.539962e+00 14.82241300
## PC16 2.054784e+01 14.15671362
## PC120 7.951659e-03 12.50922508
## PC27 6.583676e+00 12.00582798
## PC9 5.584978e+01 11.39025458
## PC66 0.000000e+00 11.03040295
## PC28 0.000000e+00 10.32604668
## PC2 1.110115e+01 10.32187287
## PC122 0.000000e+00 9.92086158
## PC139 0.000000e+00 9.77496851
## PC94 0.000000e+00 9.37136273
## PC107 0.000000e+00 9.17458911
## PC133 8.569951e-02 7.94851125
## PC62 0.000000e+00 7.09755133
## PC61 0.000000e+00 6.88128256
## PC108 6.523922e+00 6.85411474
## PC128 2.424400e+01 6.12215252
## PC106 6.709574e+00 5.36168397
## PC91 0.000000e+00 4.65591574
## PC72 0.000000e+00 3.77670850
## PC100 1.448174e+01 3.48242493
## PC71 0.000000e+00 3.43918050
## PC110 0.000000e+00 2.30904330
## PC37 2.029163e-02 2.28472436
## PC74 5.526187e+00 2.10104845
## PC20 6.471282e+00 1.77229191
## PC142 1.751430e+01 1.48251276
## PC105 0.000000e+00 1.33318291
## PC83 0.000000e+00 0.76781535
## PC43 0.000000e+00 0.58891802
## PC19 2.341047e+01 0.54601541
## PC93 0.000000e+00 0.47294227
## PC54 5.627894e+00 0.21298046
## PC55 0.000000e+00 0.09087835
## PC44 4.657604e+00 0.04517734
## PC1 1.487771e+00 0.00000000
## PC101 1.257034e+01 0.00000000
## PC102 3.345513e+00 0.00000000
## PC103 2.182427e+00 0.00000000
## PC104 0.000000e+00 0.00000000
## PC109 4.849634e+00 0.00000000
## PC11 1.060646e+01 0.00000000
## PC111 5.951208e+00 0.00000000
## PC112 0.000000e+00 0.00000000
## PC114 1.650518e+01 0.00000000
## PC115 0.000000e+00 0.00000000
## PC116 0.000000e+00 0.00000000
## PC117 5.778034e+00 0.00000000
## PC118 4.902623e+00 0.00000000
## PC119 0.000000e+00 0.00000000
## PC121 0.000000e+00 0.00000000
## PC123 4.889254e-01 0.00000000
## PC124 2.213194e+01 0.00000000
## PC125 0.000000e+00 0.00000000
## PC126 1.295801e+01 0.00000000
## PC127 0.000000e+00 0.00000000
## PC129 0.000000e+00 0.00000000
## PC13 0.000000e+00 0.00000000
## PC130 2.081387e+00 0.00000000
## PC131 0.000000e+00 0.00000000
## PC132 0.000000e+00 0.00000000
## PC134 0.000000e+00 0.00000000
## PC135 0.000000e+00 0.00000000
## PC136 0.000000e+00 0.00000000
## PC137 7.388924e-02 0.00000000
## PC138 1.321459e+01 0.00000000
## PC140 5.171006e+00 0.00000000
## PC141 0.000000e+00 0.00000000
## PC143 0.000000e+00 0.00000000
## PC17 8.207663e+00 0.00000000
## PC18 0.000000e+00 0.00000000
## PC21 0.000000e+00 0.00000000
## PC22 1.452383e+01 0.00000000
## PC24 0.000000e+00 0.00000000
## PC25 2.371905e+01 0.00000000
## PC26 5.050796e+01 0.00000000
## PC29 0.000000e+00 0.00000000
## PC30 2.569378e+01 0.00000000
## PC32 0.000000e+00 0.00000000
## PC33 1.735816e+00 0.00000000
## PC34 0.000000e+00 0.00000000
## PC35 4.402217e+00 0.00000000
## PC36 0.000000e+00 0.00000000
## PC39 0.000000e+00 0.00000000
## PC40 1.496730e+01 0.00000000
## PC41 0.000000e+00 0.00000000
## PC45 1.271547e+01 0.00000000
## PC46 2.206416e-01 0.00000000
## PC47 6.727046e+00 0.00000000
## PC48 0.000000e+00 0.00000000
## PC49 3.217595e+00 0.00000000
## PC50 0.000000e+00 0.00000000
## PC51 1.903814e+00 0.00000000
## PC52 2.128227e+01 0.00000000
## PC53 0.000000e+00 0.00000000
## PC57 3.105066e+01 0.00000000
## PC58 0.000000e+00 0.00000000
## PC59 0.000000e+00 0.00000000
## PC60 0.000000e+00 0.00000000
## PC63 0.000000e+00 0.00000000
## PC64 0.000000e+00 0.00000000
## PC65 0.000000e+00 0.00000000
## PC67 0.000000e+00 0.00000000
## PC68 2.684775e+01 0.00000000
## PC69 0.000000e+00 0.00000000
## PC70 7.297321e+00 0.00000000
## PC73 0.000000e+00 0.00000000
## PC75 1.133188e+00 0.00000000
## PC76 0.000000e+00 0.00000000
## PC78 0.000000e+00 0.00000000
## PC79 0.000000e+00 0.00000000
## PC8 1.000000e+02 0.00000000
## PC80 0.000000e+00 0.00000000
## PC81 0.000000e+00 0.00000000
## PC82 0.000000e+00 0.00000000
## PC84 3.861871e+00 0.00000000
## PC85 5.090498e+00 0.00000000
## PC87 5.856015e+00 0.00000000
## PC89 0.000000e+00 0.00000000
## PC90 0.000000e+00 0.00000000
## PC95 6.273683e+00 0.00000000
## PC96 0.000000e+00 0.00000000
## PC98 0.000000e+00 0.00000000
## PC99 3.645912e-01 0.00000000
## PC144 5.666493e+00 NA
## PC145 0.000000e+00 NA
## PC146 0.000000e+00 NA
## imp
## PC7 100.00000000
## PC10 50.56857563
## PC5 49.83764917
## PC92 43.75600909
## PC12 43.72906314
## PC3 41.98091808
## PC15 41.77916319
## PC97 29.84358506
## PC31 24.48580379
## PC77 22.60589561
## PC113 21.70310070
## PC4 20.24378933
## PC88 20.22414706
## PC56 20.21582023
## PC38 19.82002416
## PC23 19.72359791
## PC14 17.67354338
## PC42 17.34459014
## PC86 17.21752030
## PC6 14.82241300
## PC16 14.15671362
## PC120 12.50922508
## PC27 12.00582798
## PC9 11.39025458
## PC66 11.03040295
## PC28 10.32604668
## PC2 10.32187287
## PC122 9.92086158
## PC139 9.77496851
## PC94 9.37136273
## PC107 9.17458911
## PC133 7.94851125
## PC62 7.09755133
## PC61 6.88128256
## PC108 6.85411474
## PC128 6.12215252
## PC106 5.36168397
## PC91 4.65591574
## PC72 3.77670850
## PC100 3.48242493
## PC71 3.43918050
## PC110 2.30904330
## PC37 2.28472436
## PC74 2.10104845
## PC20 1.77229191
## PC142 1.48251276
## PC105 1.33318291
## PC83 0.76781535
## PC43 0.58891802
## PC19 0.54601541
## PC93 0.47294227
## PC54 0.21298046
## PC55 0.09087835
## PC44 0.04517734
## PC1 0.00000000
## PC101 0.00000000
## PC102 0.00000000
## PC103 0.00000000
## PC104 0.00000000
## PC109 0.00000000
## PC11 0.00000000
## PC111 0.00000000
## PC112 0.00000000
## PC114 0.00000000
## PC115 0.00000000
## PC116 0.00000000
## PC117 0.00000000
## PC118 0.00000000
## PC119 0.00000000
## PC121 0.00000000
## PC123 0.00000000
## PC124 0.00000000
## PC125 0.00000000
## PC126 0.00000000
## PC127 0.00000000
## PC129 0.00000000
## PC13 0.00000000
## PC130 0.00000000
## PC131 0.00000000
## PC132 0.00000000
## PC134 0.00000000
## PC135 0.00000000
## PC136 0.00000000
## PC137 0.00000000
## PC138 0.00000000
## PC140 0.00000000
## PC141 0.00000000
## PC143 0.00000000
## PC17 0.00000000
## PC18 0.00000000
## PC21 0.00000000
## PC22 0.00000000
## PC24 0.00000000
## PC25 0.00000000
## PC26 0.00000000
## PC29 0.00000000
## PC30 0.00000000
## PC32 0.00000000
## PC33 0.00000000
## PC34 0.00000000
## PC35 0.00000000
## PC36 0.00000000
## PC39 0.00000000
## PC40 0.00000000
## PC41 0.00000000
## PC45 0.00000000
## PC46 0.00000000
## PC47 0.00000000
## PC48 0.00000000
## PC49 0.00000000
## PC50 0.00000000
## PC51 0.00000000
## PC52 0.00000000
## PC53 0.00000000
## PC57 0.00000000
## PC58 0.00000000
## PC59 0.00000000
## PC60 0.00000000
## PC63 0.00000000
## PC64 0.00000000
## PC65 0.00000000
## PC67 0.00000000
## PC68 0.00000000
## PC69 0.00000000
## PC70 0.00000000
## PC73 0.00000000
## PC75 0.00000000
## PC76 0.00000000
## PC78 0.00000000
## PC79 0.00000000
## PC8 0.00000000
## PC80 0.00000000
## PC81 0.00000000
## PC82 0.00000000
## PC84 0.00000000
## PC85 0.00000000
## PC87 0.00000000
## PC89 0.00000000
## PC90 0.00000000
## PC95 0.00000000
## PC96 0.00000000
## PC98 0.00000000
## PC99 0.00000000
## PC144 NA
## PC145 NA
## PC146 NAif (glb_is_classification && glb_is_binomial)
glb_analytics_diag_plots(obs_df=glbObsTrn, mdl_id=glbMdlFinId,
prob_threshold=glb_models_df[glb_models_df$id == glbMdlSelId,
"opt.prob.threshold.OOB"]) else
glb_analytics_diag_plots(obs_df=glbObsTrn, mdl_id=glbMdlFinId) ## Warning in glb_analytics_diag_plots(obs_df = glbObsTrn, mdl_id =
## glbMdlFinId, : Limiting important feature scatter plots to 5 out of 143
## [1] "Min/Max Boundaries: "
## USER_ID Party.fctr Party.fctr.RFE.X.zv.pca.rcv.glmnet.prob
## 1 1397 R NA
## 2 3075 R NA
## 3 6112 R 0.4561332
## 4 118 R 0.7204766
## Party.fctr.RFE.X.zv.pca.rcv.glmnet
## 1 <NA>
## 2 <NA>
## 3 D
## 4 R
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.err
## 1 NA
## 2 NA
## 3 TRUE
## 4 FALSE
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.err.abs
## 1 NA
## 2 NA
## 3 0.5438668
## 4 0.2795234
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.is.acc
## 1 NA
## 2 NA
## 3 FALSE
## 4 TRUE
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.prob
## 1 0.1366892
## 2 0.3074388
## 3 0.4867866
## 4 0.6803799
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet
## 1 D
## 2 D
## 3 D
## 4 R
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.err
## 1 TRUE
## 2 TRUE
## 3 TRUE
## 4 FALSE
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.err.abs
## 1 0.8633108
## 2 0.6925612
## 3 0.5132134
## 4 0.3196201
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.is.acc
## 1 FALSE
## 2 FALSE
## 3 FALSE
## 4 TRUE
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.accurate
## 1 FALSE
## 2 FALSE
## 3 FALSE
## 4 TRUE
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.error .label
## 1 -0.36331085 1397
## 2 -0.19256116 3075
## 3 -0.01321344 6112
## 4 0.00000000 118
## [1] "Inaccurate: "
## USER_ID Party.fctr Party.fctr.RFE.X.zv.pca.rcv.glmnet.prob
## 1 3895 R NA
## 2 626 R 0.07256962
## 3 2226 R 0.10893408
## 4 2590 R 0.09674309
## 5 1397 R NA
## 6 468 R NA
## Party.fctr.RFE.X.zv.pca.rcv.glmnet
## 1 <NA>
## 2 D
## 3 D
## 4 D
## 5 <NA>
## 6 <NA>
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.err
## 1 NA
## 2 TRUE
## 3 TRUE
## 4 TRUE
## 5 NA
## 6 NA
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.err.abs
## 1 NA
## 2 0.9274304
## 3 0.8910659
## 4 0.9032569
## 5 NA
## 6 NA
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.is.acc
## 1 NA
## 2 FALSE
## 3 FALSE
## 4 FALSE
## 5 NA
## 6 NA
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.prob
## 1 0.1081122
## 2 0.1194254
## 3 0.1251642
## 4 0.1347899
## 5 0.1366892
## 6 0.1406591
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet
## 1 D
## 2 D
## 3 D
## 4 D
## 5 D
## 6 D
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.err
## 1 TRUE
## 2 TRUE
## 3 TRUE
## 4 TRUE
## 5 TRUE
## 6 TRUE
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.err.abs
## 1 0.8918878
## 2 0.8805746
## 3 0.8748358
## 4 0.8652101
## 5 0.8633108
## 6 0.8593409
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.is.acc
## 1 FALSE
## 2 FALSE
## 3 FALSE
## 4 FALSE
## 5 FALSE
## 6 FALSE
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.accurate
## 1 FALSE
## 2 FALSE
## 3 FALSE
## 4 FALSE
## 5 FALSE
## 6 FALSE
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.error
## 1 -0.3918878
## 2 -0.3805746
## 3 -0.3748358
## 4 -0.3652101
## 5 -0.3633108
## 6 -0.3593409
## USER_ID Party.fctr Party.fctr.RFE.X.zv.pca.rcv.glmnet.prob
## 644 2468 R 0.4491806
## 844 2744 R NA
## 1150 4742 D NA
## 1369 2429 D 0.6985087
## 1586 5400 D 0.5482780
## 1828 3771 D 0.6357115
## Party.fctr.RFE.X.zv.pca.rcv.glmnet
## 644 D
## 844 <NA>
## 1150 <NA>
## 1369 R
## 1586 R
## 1828 R
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.err
## 644 TRUE
## 844 NA
## 1150 NA
## 1369 TRUE
## 1586 TRUE
## 1828 TRUE
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.err.abs
## 644 0.5508194
## 844 NA
## 1150 NA
## 1369 0.6985087
## 1586 0.5482780
## 1828 0.6357115
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.is.acc
## 644 FALSE
## 844 NA
## 1150 NA
## 1369 FALSE
## 1586 FALSE
## 1828 FALSE
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.prob
## 644 0.4460571
## 844 0.4712180
## 1150 0.5045288
## 1369 0.5354487
## 1586 0.5725441
## 1828 0.6606227
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet
## 644 D
## 844 D
## 1150 R
## 1369 R
## 1586 R
## 1828 R
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.err
## 644 TRUE
## 844 TRUE
## 1150 TRUE
## 1369 TRUE
## 1586 TRUE
## 1828 TRUE
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.err.abs
## 644 0.5539429
## 844 0.5287820
## 1150 0.5045288
## 1369 0.5354487
## 1586 0.5725441
## 1828 0.6606227
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.is.acc
## 644 FALSE
## 844 FALSE
## 1150 FALSE
## 1369 FALSE
## 1586 FALSE
## 1828 FALSE
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.accurate
## 644 FALSE
## 844 FALSE
## 1150 FALSE
## 1369 FALSE
## 1586 FALSE
## 1828 FALSE
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.error
## 644 -0.053942908
## 844 -0.028782008
## 1150 0.004528837
## 1369 0.035448749
## 1586 0.072544081
## 1828 0.160622724
## USER_ID Party.fctr Party.fctr.RFE.X.zv.pca.rcv.glmnet.prob
## 1933 5889 D 0.8712831
## 1934 1064 D 0.8388104
## 1935 3374 D 0.8388935
## 1936 1309 D NA
## 1937 217 D NA
## 1938 3433 D 0.8011552
## Party.fctr.RFE.X.zv.pca.rcv.glmnet
## 1933 R
## 1934 R
## 1935 R
## 1936 <NA>
## 1937 <NA>
## 1938 R
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.err
## 1933 TRUE
## 1934 TRUE
## 1935 TRUE
## 1936 NA
## 1937 NA
## 1938 TRUE
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.err.abs
## 1933 0.8712831
## 1934 0.8388104
## 1935 0.8388935
## 1936 NA
## 1937 NA
## 1938 0.8011552
## Party.fctr.RFE.X.zv.pca.rcv.glmnet.is.acc
## 1933 FALSE
## 1934 FALSE
## 1935 FALSE
## 1936 NA
## 1937 NA
## 1938 FALSE
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.prob
## 1933 0.7958976
## 1934 0.8017442
## 1935 0.8060458
## 1936 0.8156517
## 1937 0.8327223
## 1938 0.8519443
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet
## 1933 R
## 1934 R
## 1935 R
## 1936 R
## 1937 R
## 1938 R
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.err
## 1933 TRUE
## 1934 TRUE
## 1935 TRUE
## 1936 TRUE
## 1937 TRUE
## 1938 TRUE
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.err.abs
## 1933 0.7958976
## 1934 0.8017442
## 1935 0.8060458
## 1936 0.8156517
## 1937 0.8327223
## 1938 0.8519443
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.is.acc
## 1933 FALSE
## 1934 FALSE
## 1935 FALSE
## 1936 FALSE
## 1937 FALSE
## 1938 FALSE
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.accurate
## 1933 FALSE
## 1934 FALSE
## 1935 FALSE
## 1936 FALSE
## 1937 FALSE
## 1938 FALSE
## Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.error
## 1933 0.2958976
## 1934 0.3017442
## 1935 0.3060458
## 1936 0.3156517
## 1937 0.3327223
## 1938 0.3519443
dsp_feats_vctr <- c(NULL)
for(var in grep(".imp", names(glb_feats_df), fixed=TRUE, value=TRUE))
dsp_feats_vctr <- union(dsp_feats_vctr,
glb_feats_df[!is.na(glb_feats_df[, var]), "id"])
# print(glbObsTrn[glbObsTrn$UniqueID %in% FN_OOB_ids,
# grep(glb_rsp_var, names(glbObsTrn), value=TRUE)])
print(setdiff(names(glbObsTrn), names(glbObsAll)))## [1] "Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.prob"
## [2] "Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet"
## [3] "Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.err"
## [4] "Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.err.abs"
## [5] "Party.fctr.Final.RFE.X.zv.pca.rcv.glmnet.is.acc"for (col in setdiff(names(glbObsTrn), names(glbObsAll)))
# Merge or cbind ?
glbObsAll[glbObsAll$.src == "Train", col] <- glbObsTrn[, col]
print(setdiff(names(glbObsFit), names(glbObsAll)))## character(0)print(setdiff(names(glbObsOOB), names(glbObsAll)))## character(0)for (col in setdiff(names(glbObsOOB), names(glbObsAll)))
# Merge or cbind ?
glbObsAll[glbObsAll$.lcn == "OOB", col] <- glbObsOOB[, col]
print(setdiff(names(glbObsNew), names(glbObsAll)))## character(0)#glb2Sav(); all.equal(savObsAll, glbObsAll); all.equal(sav_models_lst, glb_models_lst)
#load(file = paste0(glbOut$pfx, "dsk_knitr.RData"))
#cmpCols <- names(glbObsAll)[!grepl("\\.Final\\.", names(glbObsAll))]; all.equal(savObsAll[, cmpCols], glbObsAll[, cmpCols]); all.equal(savObsAll[, "H.P.http"], glbObsAll[, "H.P.http"]);
replay.petrisim(pn = glb_analytics_pn,
replay.trans = (glb_analytics_avl_objs <- c(glb_analytics_avl_objs,
"data.training.all.prediction","model.final")), flip_coord = TRUE)## time trans "bgn " "fit.data.training.all " "predict.data.new " "end "
## 0.0000 multiple enabled transitions: data.training.all data.new model.selected firing: model.selected
## 1.0000 3 2 1 0 0
## 1.0000 multiple enabled transitions: data.training.all data.new model.selected model.final data.training.all.prediction firing: data.training.all.prediction
## 2.0000 5 2 0 0 1## Warning in replay.petrisim(pn = glb_analytics_pn, replay.trans =
## (glb_analytics_avl_objs <- c(glb_analytics_avl_objs, : Transition:
## model.final not enabled; adding missing token(s)## Warning in replay.petrisim(pn = glb_analytics_pn, replay.trans
## = (glb_analytics_avl_objs <- c(glb_analytics_avl_objs, : Place:
## fit.data.training.all: added 1 missing token## 2.0000 multiple enabled transitions: data.training.all data.new model.selected model.final data.training.all.prediction firing: model.final
## 3.0000 4 2 0 1 1
glb_chunks_df <- myadd_chunk(glb_chunks_df, "predict.data.new", major.inc = TRUE)## label step_major step_minor label_minor bgn end
## 5 fit.data.training 2 1 1 517.208 529.463
## 6 predict.data.new 3 0 0 529.464 NA
## elapsed
## 5 12.255
## 6 NAStep 3.0: predict data new
## Warning in glb_get_predictions(obs_df, mdl_id = glbMdlFinId, rsp_var =
## glb_rsp_var, : Using default probability threshold: 0.5
## Warning in glb_get_predictions(obs_df, mdl_id = glbMdlFinId, rsp_var =
## glb_rsp_var, : Using default probability threshold: 0.5## Warning in glb_analytics_diag_plots(obs_df = glbObsNew, mdl_id =
## glbMdlFinId, : Limiting important feature scatter plots to 5 out of 143## Warning: Removed 1392 rows containing missing values (geom_point).
## Warning: Removed 1392 rows containing missing values (geom_point).
## Warning: Removed 1392 rows containing missing values (geom_point).
## Warning: Removed 1392 rows containing missing values (geom_point).
## Warning: Removed 1392 rows containing missing values (geom_point).
## Warning: Removed 1392 rows containing missing values (geom_point).
## Warning: Removed 1392 rows containing missing values (geom_point).
## Warning: Removed 1392 rows containing missing values (geom_point).
## Warning: Removed 1392 rows containing missing values (geom_point).
## Warning: Removed 1392 rows containing missing values (geom_point).
## NULL## Loading required package: tidyr##
## Attaching package: 'tidyr'## The following object is masked from 'package:Matrix':
##
## expand## [1] "OOBobs total range outliers: 0"
## [1] "newobs total range outliers: 0"## [1] 0.5## [1] "glbMdlSelId: RFE.X#zv.pca#rcv#glmnet"## [1] "glbMdlFinId: Final.RFE.X#zv.pca#rcv#glmnet"## [1] "Cross Validation issues:"
## MFO###myMFO_classfr Random###myrandom_classfr
## 0 0
## Max.cor.Y.rcv.1X1###glmnet
## 0## max.Accuracy.OOB max.AUCROCR.OOB
## RFE.X#zv.pca#rcv#glmnet 0.5829596 0.5816510
## All.X#nzv.spatialSign#rcv#glmnet 0.5820628 0.5787416
## All.X#spatialSign#rcv#glmnet 0.5811659 0.5793099
## RFE.X#nzv.spatialSign#rcv#glmnet 0.5811659 0.5785672
## All.X.Inc#nzv.spatialSign#rcv#glmnet 0.5811659 0.5734523
## RFE.X#spatialSign#rcv#glmnet 0.5802691 0.5794035
## All.X#nzv#rcv#glmnet 0.5775785 0.5766539
## RFE.X#nzv#rcv#glmnet 0.5775785 0.5762019
## All.X#expoTrans#rcv#glmnet 0.5775785 0.5745421
## RFE.X#expoTrans#rcv#glmnet 0.5775785 0.5745421
## RFE.X#YeoJohnson#rcv#glmnet 0.5775785 0.5744824
## All.X#YeoJohnson#rcv#glmnet 0.5775785 0.5744824
## RFE.X##rcv#glmnet 0.5766816 0.5757966
## RFE.X#zv#rcv#glmnet 0.5766816 0.5757966
## RFE.X#BoxCox#rcv#glmnet 0.5766816 0.5757966
## RFE.X#center#rcv#glmnet 0.5766816 0.5757966
## RFE.X#scale#rcv#glmnet 0.5766816 0.5757966
## RFE.X#center.scale#rcv#glmnet 0.5766816 0.5757966
## RFE.X#range#rcv#glmnet 0.5766816 0.5757966
## RFE.X#conditionalX#rcv#glmnet 0.5766816 0.5757966
## Low.cor.X##rcv#glmnet 0.5766816 0.5757966
## All.X##rcv#glmnet 0.5766816 0.5757966
## All.X#zv#rcv#glmnet 0.5766816 0.5757966
## All.X#BoxCox#rcv#glmnet 0.5766816 0.5757966
## All.X#center#rcv#glmnet 0.5766816 0.5757966
## All.X#scale#rcv#glmnet 0.5766816 0.5757966
## All.X#center.scale#rcv#glmnet 0.5766816 0.5757966
## All.X#range#rcv#glmnet 0.5766816 0.5757966
## All.X#conditionalX#rcv#glmnet 0.5766816 0.5757966
## All.X#zv.pca#rcv#glmnet 0.5748879 0.5798104
## RFE.X#nzv.pca.spatialSign#rcv#glmnet 0.5713004 0.5742305
## All.X##rcv#glm 0.5650224 0.5687249
## RFE.X#ica#rcv#glmnet 0.5363229 0.5370894
## All.X#ica#rcv#glmnet 0.5354260 0.5360626
## Interact.High.cor.Y##rcv#glmnet 0.5345291 0.5242069
## Random###myrandom_classfr 0.5300448 0.5181895
## Max.cor.Y.rcv.1X1###glmnet 0.5300448 0.5102459
## Max.cor.Y##rcv#rpart 0.5300448 0.5000646
## MFO###myMFO_classfr 0.5300448 0.5000000
## Final.RFE.X#zv.pca#rcv#glmnet NA NA
## max.AUCpROC.OOB max.Accuracy.fit
## RFE.X#zv.pca#rcv#glmnet 0.5792308 0.6459313
## All.X#nzv.spatialSign#rcv#glmnet 0.5572148 0.6478785
## All.X#spatialSign#rcv#glmnet 0.5570372 0.6492270
## RFE.X#nzv.spatialSign#rcv#glmnet 0.5597528 0.6462317
## All.X.Inc#nzv.spatialSign#rcv#glmnet 0.5556503 0.6440612
## RFE.X#spatialSign#rcv#glmnet 0.5570372 0.6490772
## All.X#nzv#rcv#glmnet 0.5550900 0.6475806
## RFE.X#nzv#rcv#glmnet 0.5531816 0.6466826
## All.X#expoTrans#rcv#glmnet 0.5559554 0.6463076
## RFE.X#expoTrans#rcv#glmnet 0.5559554 0.6462327
## RFE.X#YeoJohnson#rcv#glmnet 0.5559554 0.6466820
## All.X#YeoJohnson#rcv#glmnet 0.5559554 0.6465321
## RFE.X##rcv#glmnet 0.5588180 0.6475052
## RFE.X#zv#rcv#glmnet 0.5588180 0.6475052
## RFE.X#BoxCox#rcv#glmnet 0.5588180 0.6475052
## RFE.X#center#rcv#glmnet 0.5588180 0.6475052
## RFE.X#scale#rcv#glmnet 0.5588180 0.6475052
## RFE.X#center.scale#rcv#glmnet 0.5588180 0.6475052
## RFE.X#range#rcv#glmnet 0.5588180 0.6475052
## RFE.X#conditionalX#rcv#glmnet 0.5588180 0.6475052
## Low.cor.X##rcv#glmnet 0.5588180 0.6471308
## All.X##rcv#glmnet 0.5588180 0.6471308
## All.X#zv#rcv#glmnet 0.5588180 0.6471308
## All.X#BoxCox#rcv#glmnet 0.5588180 0.6471308
## All.X#center#rcv#glmnet 0.5588180 0.6471308
## All.X#scale#rcv#glmnet 0.5588180 0.6471308
## All.X#center.scale#rcv#glmnet 0.5588180 0.6471308
## All.X#range#rcv#glmnet 0.5588180 0.6471308
## All.X#conditionalX#rcv#glmnet 0.5588180 0.6471308
## All.X#zv.pca#rcv#glmnet 0.5717247 0.6421158
## RFE.X#nzv.pca.spatialSign#rcv#glmnet 0.5545298 0.6480297
## All.X##rcv#glm 0.5487933 0.6254219
## RFE.X#ica#rcv#glmnet 0.5121737 0.5603719
## All.X#ica#rcv#glmnet 0.5133442 0.5607462
## Interact.High.cor.Y##rcv#glmnet 0.5218093 0.6250526
## Random###myrandom_classfr 0.4836608 0.5299798
## Max.cor.Y.rcv.1X1###glmnet 0.4999322 0.6240737
## Max.cor.Y##rcv#rpart 0.4999322 0.6227308
## MFO###myMFO_classfr 0.5000000 0.5299798
## Final.RFE.X#zv.pca#rcv#glmnet NA 0.6272737
## opt.prob.threshold.fit
## RFE.X#zv.pca#rcv#glmnet 0.50
## All.X#nzv.spatialSign#rcv#glmnet 0.50
## All.X#spatialSign#rcv#glmnet 0.50
## RFE.X#nzv.spatialSign#rcv#glmnet 0.50
## All.X.Inc#nzv.spatialSign#rcv#glmnet 0.50
## RFE.X#spatialSign#rcv#glmnet 0.50
## All.X#nzv#rcv#glmnet 0.50
## RFE.X#nzv#rcv#glmnet 0.50
## All.X#expoTrans#rcv#glmnet 0.50
## RFE.X#expoTrans#rcv#glmnet 0.50
## RFE.X#YeoJohnson#rcv#glmnet 0.50
## All.X#YeoJohnson#rcv#glmnet 0.50
## RFE.X##rcv#glmnet 0.50
## RFE.X#zv#rcv#glmnet 0.50
## RFE.X#BoxCox#rcv#glmnet 0.50
## RFE.X#center#rcv#glmnet 0.50
## RFE.X#scale#rcv#glmnet 0.50
## RFE.X#center.scale#rcv#glmnet 0.50
## RFE.X#range#rcv#glmnet 0.50
## RFE.X#conditionalX#rcv#glmnet 0.50
## Low.cor.X##rcv#glmnet 0.50
## All.X##rcv#glmnet 0.50
## All.X#zv#rcv#glmnet 0.50
## All.X#BoxCox#rcv#glmnet 0.50
## All.X#center#rcv#glmnet 0.50
## All.X#scale#rcv#glmnet 0.50
## All.X#center.scale#rcv#glmnet 0.50
## All.X#range#rcv#glmnet 0.50
## All.X#conditionalX#rcv#glmnet 0.50
## All.X#zv.pca#rcv#glmnet 0.50
## RFE.X#nzv.pca.spatialSign#rcv#glmnet 0.50
## All.X##rcv#glm 0.50
## RFE.X#ica#rcv#glmnet 0.50
## All.X#ica#rcv#glmnet 0.50
## Interact.High.cor.Y##rcv#glmnet 0.50
## Random###myrandom_classfr 0.55
## Max.cor.Y.rcv.1X1###glmnet 0.45
## Max.cor.Y##rcv#rpart 0.50
## MFO###myMFO_classfr 0.50
## Final.RFE.X#zv.pca#rcv#glmnet 0.50
## opt.prob.threshold.OOB
## RFE.X#zv.pca#rcv#glmnet 0.50
## All.X#nzv.spatialSign#rcv#glmnet 0.60
## All.X#spatialSign#rcv#glmnet 0.60
## RFE.X#nzv.spatialSign#rcv#glmnet 0.60
## All.X.Inc#nzv.spatialSign#rcv#glmnet 0.60
## RFE.X#spatialSign#rcv#glmnet 0.60
## All.X#nzv#rcv#glmnet 0.60
## RFE.X#nzv#rcv#glmnet 0.55
## All.X#expoTrans#rcv#glmnet 0.55
## RFE.X#expoTrans#rcv#glmnet 0.55
## RFE.X#YeoJohnson#rcv#glmnet 0.55
## All.X#YeoJohnson#rcv#glmnet 0.55
## RFE.X##rcv#glmnet 0.55
## RFE.X#zv#rcv#glmnet 0.55
## RFE.X#BoxCox#rcv#glmnet 0.55
## RFE.X#center#rcv#glmnet 0.55
## RFE.X#scale#rcv#glmnet 0.55
## RFE.X#center.scale#rcv#glmnet 0.55
## RFE.X#range#rcv#glmnet 0.55
## RFE.X#conditionalX#rcv#glmnet 0.55
## Low.cor.X##rcv#glmnet 0.55
## All.X##rcv#glmnet 0.55
## All.X#zv#rcv#glmnet 0.55
## All.X#BoxCox#rcv#glmnet 0.55
## All.X#center#rcv#glmnet 0.55
## All.X#scale#rcv#glmnet 0.55
## All.X#center.scale#rcv#glmnet 0.55
## All.X#range#rcv#glmnet 0.55
## All.X#conditionalX#rcv#glmnet 0.55
## All.X#zv.pca#rcv#glmnet 0.50
## RFE.X#nzv.pca.spatialSign#rcv#glmnet 0.55
## All.X##rcv#glm 0.65
## RFE.X#ica#rcv#glmnet 0.55
## All.X#ica#rcv#glmnet 0.55
## Interact.High.cor.Y##rcv#glmnet 0.70
## Random###myrandom_classfr 0.55
## Max.cor.Y.rcv.1X1###glmnet 0.65
## Max.cor.Y##rcv#rpart 0.65
## MFO###myMFO_classfr 0.50
## Final.RFE.X#zv.pca#rcv#glmnet NA## [1] "RFE.X#zv.pca#rcv#glmnet OOB confusion matrix & accuracy: "
## Prediction
## Reference D R
## D 379 212
## R 253 271## err.abs.fit.sum err.abs.OOB.sum err.abs.trn.sum err.abs.new.sum
## No 883.9618 239.90734 1138.9447 NA
## NA 821.7109 209.08425 1036.9025 NA
## Yes 200.4801 85.28436 335.1653 NA
## .freqRatio.Fit .freqRatio.OOB .freqRatio.Tst .n.Fit .n.New.D .n.New.R
## No 0.4403773 0.4466368 0.4468391 1961 223 399
## NA 0.3920952 0.3928251 0.3929598 1746 357 190
## Yes 0.1675275 0.1605381 0.1602011 746 216 7
## .n.OOB .n.Trn.D .n.Trn.R .n.Tst .n.fit .n.new .n.trn err.abs.OOB.mean
## No 498 1038 1421 622 1961 622 2459 0.4817416
## NA 438 1171 1013 547 1746 547 2184 0.4773613
## Yes 179 742 183 223 746 223 925 0.4764489
## err.abs.fit.mean err.abs.new.mean err.abs.trn.mean
## No 0.4507709 NA 0.4631739
## NA 0.4706248 NA 0.4747722
## Yes 0.2687400 NA 0.3623408
## err.abs.fit.sum err.abs.OOB.sum err.abs.trn.sum err.abs.new.sum
## 1906.152762 534.275955 2511.012416 NA
## .freqRatio.Fit .freqRatio.OOB .freqRatio.Tst .n.Fit
## 1.000000 1.000000 1.000000 4453.000000
## .n.New.D .n.New.R .n.OOB .n.Trn.D
## 796.000000 596.000000 1115.000000 2951.000000
## .n.Trn.R .n.Tst .n.fit .n.new
## 2617.000000 1392.000000 4453.000000 1392.000000
## .n.trn err.abs.OOB.mean err.abs.fit.mean err.abs.new.mean
## 5568.000000 1.435552 1.190136 NA
## err.abs.trn.mean
## 1.300287## [1] "Features Importance for selected models:"## RFE.X.zv.pca.rcv.glmnet.imp Final.RFE.X.zv.pca.rcv.glmnet.imp
## PC8 1.000000e+02 0.0000000
## PC5 7.995721e+01 49.8376492
## PC10 6.660214e+01 50.5685756
## PC15 6.321726e+01 41.7791632
## PC9 5.584978e+01 11.3902546
## PC26 5.050796e+01 0.0000000
## PC3 4.474537e+01 41.9809181
## PC14 4.398423e+01 17.6735434
## PC57 3.105066e+01 0.0000000
## PC68 2.684775e+01 0.0000000
## PC30 2.569378e+01 0.0000000
## PC88 2.477208e+01 20.2241471
## PC128 2.424400e+01 6.1221525
## PC25 2.371905e+01 0.0000000
## PC19 2.341047e+01 0.5460154
## PC124 2.213194e+01 0.0000000
## PC52 2.128227e+01 0.0000000
## PC16 2.054784e+01 14.1567136
## PC142 1.751430e+01 1.4825128
## PC114 1.650518e+01 0.0000000
## PC40 1.496730e+01 0.0000000
## PC23 1.488607e+01 19.7235979
## PC22 1.452383e+01 0.0000000
## PC100 1.448174e+01 3.4824249
## PC138 1.321459e+01 0.0000000
## PC126 1.295801e+01 0.0000000
## PC45 1.271547e+01 0.0000000
## PC101 1.257034e+01 0.0000000
## PC2 1.110115e+01 10.3218729
## PC11 1.060646e+01 0.0000000
## PC77 8.954286e+00 22.6058956
## PC6 8.539962e+00 14.8224130
## PC27 6.583676e+00 12.0058280
## PC92 6.234515e+00 43.7560091
## PC4 4.495763e+00 20.2437893
## PC113 2.752156e+00 21.7031007
## PC12 1.279928e-01 43.7290631
## PC120 7.951659e-03 12.5092251
## PC7 0.000000e+00 100.0000000
## PC97 0.000000e+00 29.8435851
## PC31 0.000000e+00 24.4858038
## PC56 0.000000e+00 20.2158202
## PC38 0.000000e+00 19.8200242
## PC42 0.000000e+00 17.3445901
## PC86 0.000000e+00 17.2175203
## PC66 0.000000e+00 11.0304029
## PC28 0.000000e+00 10.3260467## [1] "glbObsNew prediction stats:"##
## D R
## 796 596## label step_major step_minor label_minor bgn end
## 6 predict.data.new 3 0 0 529.464 573.709
## 7 display.session.info 4 0 0 573.710 NA
## elapsed
## 6 44.245
## 7 NANull Hypothesis (\(\sf{H_{0}}\)): mpg is not impacted by am_fctr.
The variance by am_fctr appears to be independent. #{r q1, cache=FALSE} # print(t.test(subset(cars_df, am_fctr == "automatic")$mpg, # subset(cars_df, am_fctr == "manual")$mpg, # var.equal=FALSE)$conf) # We reject the null hypothesis i.e. we have evidence to conclude that am_fctr impacts mpg (95% confidence). Manual transmission is better for miles per gallon versus automatic transmission.
## label step_major step_minor label_minor bgn end
## 1 fit.models_1 1 0 0 7.720 319.664
## 4 fit.data.training 2 0 0 339.957 517.207
## 6 predict.data.new 3 0 0 529.464 573.709
## 2 fit.models 1 1 1 319.665 334.110
## 5 fit.data.training 2 1 1 517.208 529.463
## 3 fit.models 1 2 2 334.111 339.957
## elapsed duration
## 1 311.944 311.944
## 4 177.251 177.250
## 6 44.245 44.245
## 2 14.445 14.445
## 5 12.255 12.255
## 3 5.846 5.846
## [1] "Total Elapsed Time: 573.709 secs"
## label step_major step_minor label_minor bgn end
## 9 fit.models_1_preProc 1 8 preProc 318.316 319.658
## 5 fit.models_1_All.X.Inc 1 4 setup 317.999 318.291
## 1 fit.models_1_bgn 1 0 setup 317.965 317.978
## 2 fit.models_1_All.X 1 1 setup 317.978 317.988
## 6 fit.models_1_All.X.Inc 1 5 glmnet 318.292 318.301
## 8 fit.models_1_RFE.X 1 7 glmnet 318.308 318.316
## 7 fit.models_1_RFE.X 1 6 setup 318.302 318.308
## 3 fit.models_1_All.X 1 2 glmnet 317.988 317.993
## 4 fit.models_1_All.X 1 3 glm 317.994 317.998
## elapsed duration
## 9 1.342 1.342
## 5 0.292 0.292
## 1 0.013 0.013
## 2 0.010 0.010
## 6 0.009 0.009
## 8 0.008 0.008
## 7 0.006 0.006
## 3 0.005 0.005
## 4 0.004 0.004
## [1] "Total Elapsed Time: 319.658 secs"